[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7A2DFPZ3SZ1GD02A8HSDTQ2":3,"skill-same:01M7A2DFPZ3SZ1GD02A8HSDTQ2":51},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M7A2DFPZ3SZ1GD02A8HSDTQ2","author-brief",null,"Author and structure a phase plan as self-contained briefs organized into dependency waves, with a critical path and mappings to delivery milestones. Use this WHENEVER the user wants to plan multi-step work and split it up: \"break this into briefs and waves\", \"create a phase plan\", \"do for phase N what we did for phase 1\", \"this brief is too big, split it\", \"turn this into tasks with dependencies\", or when scoping a body of work that several people\u002Fagents will pick up. Encodes a brief format (Wave \u002F Depends-on \u002F Unblocks \u002F Effort \u002F Deliverables \u002F DoD), a phase-README status-table + critical-path + dependency-wave layout, and the lesson that you must verify the REAL blocker at the head of the critical path rather than assume it. Prefer this over inventing an ad-hoc plan shape.","medici-finance\u002Fassay","medici-finance","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F298844171?v=4","01M7A2655EZBRV4HA62GMGC7KK","Medici","01M7A26562DY5FSYQMXSHM5EJ6",7,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Fplugins\u002Fassay\u002Fskills\u002Fauthor-brief\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: author-brief\ndescription: >-\n  Author and structure a phase plan as self-contained briefs organized into dependency\n  waves, with a critical path and mappings to delivery milestones. Use this WHENEVER the user wants to plan multi-step work and split it\n  up: \"break this into briefs and waves\", \"create a phase plan\", \"do for phase N what we did for\n  phase 1\", \"this brief is too big, split it\", \"turn this into tasks with dependencies\", or when\n  scoping a body of work that several people\u002Fagents will pick up. Encodes a brief format\n  (Wave \u002F Depends-on \u002F Unblocks \u002F Effort \u002F Deliverables \u002F DoD), a phase-README status-table +\n  critical-path + dependency-wave layout, and the lesson that you must verify the REAL blocker at the\n  head of the critical path rather than assume it. Prefer this over inventing an ad-hoc plan shape.\n---\n\n# Author Brief\n\n**A stream board's Briefs table is a derived, generated surface once `statusgen` is wired up** — a\nbrief states the authoring facts and the edges\n(`why:`, `depends`, `unblocks`, `gate`, `risk`); it never states the lifecycle cell, and no one\nhand-edits the generated table.\n\n## Model-tier gate (check BEFORE anything else)\n\nBrief authoring and work decomposition are **design-tier work**: this is where risk gates are\nderived, dependencies mapped, and the critical-path head verified — errors made here COMPOUND\nthrough every downstream implementer and review cycle, unlike implementation errors, which gates\ncatch. Therefore:\n\n- **If you are a fast\u002Fcheap-tier model** (a haiku-class or equivalent economy tier): **STOP.**\n  Report which model you are and tell the user to switch to a strong-tier session or escalate\n  the authoring to one. **Do not author anyway** — not for \"just a small brief set\", not as a\n  \"draft for review\", not because the user seems in a hurry. A cheap draft anchors the strong\n  model that reviews it; refusing is the cheaper path.\n- **If you are dispatching this work to a subagent**: use the most capable model available for\n  authoring\u002Fdecomposition — never an economy tier. (Implementation is the inverse: cheap\n  implementers behind strong gates.)\n\nPlanning multi-step work follows a consistent shape: a phase README that holds a **status table**, a\n**critical path**, and **dependency waves**, plus one **self-contained brief** per unit of work. The\npoint of the structure is that any single agent can pick up a brief and execute it without re-reading\nthe whole plan, and the reader can see at a glance what blocks what. **Match the phase docs already in\nthe repo — don't invent a new format.** If a project-level `author-brief` skill exists in the current\nrepo (`.claude\u002Fskills\u002Fauthor-brief\u002FSKILL.md`), read it after this skill and apply its specifics —\npaths, docs-regen commands, and local examples layer on top of this methodology.\n\n## Bind implementation pickup to current context\n\nRead the [implementation pickup procedure](..\u002F..\u002Freferences\u002Fimplementation-pickup.md).\nSpecs and briefs identify the selected target\u002Ftransition, semantic contracts and\napplicable project engineering guidelines by canonical pointer. Include their\npickup\u002Fresume assessment in the implementation handoff; record revisions, coverage\nand continue\u002Fadapt\u002Fhold rather than assume the authoring snapshot is still current.\nUse the shipped context operation where the installed version supports it (see the\nsupported version below); otherwise use the explicit manual bootstrap.\nDo not rewrite completed Evidence or let a draft proposal veto approved work.\n\n## Consume the context handoff before opening the PR\n\nWhen the spec hands over a context receipt, the brief PR answers it:\n\n1. Recheck the receipt against the current snapshot with\n   `statusgen context recheck --snapshot DIR receipt.json`. Reconcile (exit 1) or resolve the\n   hold (exit 3) before authoring on it.\n2. For each impact you dispose `amend`, write the candidate brief patch, a single-file\n   unified diff against the brief's current revision, and record the obligation in the\n   receipt's `amendments` with that patch's base revision and digest. Bind every obligation\n   in an `amendment-manifest-v1` file: owning brief, action, delivery paths, test commands,\n   active-work treatment, and the patch or follow-up brief it binds to.\n3. Run `statusgen context check --snapshot DIR receipt.json manifest.json PATCH...` and open\n   the PR only on exit 0. An `amend` obligation without a patch is a failure, not a note:\n   the receipt's own validation refuses it. Fix a refusal (exit 1) or a hold (exit 3); never\n   waive it in the PR body.\n4. The existing rules still apply: author `consumers:`, `depends:` \u002F `unblocks:` and the\n   Verify table as below. A patch never edits populated Evidence or a dated source account,\n   and a delivered brief takes a linked follow-on instead of a patch.\n5. Keep product and architecture bindings separate: carry each brief's `outcome` and\n   `outcome-revision` unchanged unless the spec moves them, and never drop the revision the\n   work was first selected against.\n\n**Supported version.** `statusgen context discover`, `statusgen context recheck` and\n`statusgen context check`, with the `context-handoff-v1` and `context-handoff-v2` receipts and\nthe `amendment-manifest-v1` file, ship in Assay v1.0.55 and later. The product-binding\nextension, `context-handoff-v3`, ships in v1.0.63 and later. `statusgen context --help`\nlists the verbs an install has.\n\n**Manual bootstrap.** When the installed statusgen lacks the verb, or lacks the\nproduct-binding extension while briefs bind `outcome-revision`, follow the\n[implementation pickup procedure](..\u002F..\u002Freferences\u002Fimplementation-pickup.md) by hand: record\nthe impacted briefs, their revisions, coverage and each amendment obligation in the handoff,\nand say that it is a manual bootstrap. Never invent a receipt or a schema version the\ninstall does not ship.\n\n## Map work to delivery milestones\n\nWhen the source defines delivery milestones or the request needs staged outcomes,\nread the [shared milestone authoring procedure](..\u002F..\u002Freferences\u002Fmilestone-authoring.md),\nincluding its support check and **Author-brief** section. Map existing\u002Fnew briefs to\n`\u003Cstream>:m-\u003CNN>` outcomes and criteria; prefer `\u003Cstream>:b-\u003CNN>` for new briefs\nwith optional `@\u003Cbrief-version>` on references; use full\n`\u003Ccell>:\u003Crepo>:\u003Cstream>:b-\u003CNN>` IDs in brief definitions. Once the pinned generator\nand required readers support brief §3.8, write colon-only forms and preserve legacy\naliases as compatibility reads. Until then use\nsupported legacy IDs for executable briefs and `b-` labels in scoping proposals;\ndisclose the limitation. Do not rewrite existing identities or Evidence. Membership is\nowned by the milestone, not a second brief field. Separate actual brief start needs\nfrom prerequisites for final outcome acceptance. If the source lacks a consequential\noutcome\u002Fscope decision, reconcile it with `author-spec` before committing affected work;\ncontinue independent authoring. Unsupported declarations remain proposals, not live YAML.\n\nFor a source using simulated dependencies, apply the shared procedure's **Outcomes\ndelivered with simulated dependencies** section. Preserve its real\u002Fsubstitute matrix\nand evidence limits in the affected briefs. Separate profile setup, contract\u002Ffake work\nand the joined real-consumer qualification where they have different owners; record\nactual producer prerequisites rather than adding another controller to hide a gap.\n\n## The most important judgment call: find the REAL head of the critical path\n\nThe critical path is the chain that, if it slips, slips everything. The recurring failure is putting\nthe wrong item at the head — sequencing work behind a step that is itself blocked. (Concrete shape of\nthe failure: a plan listed Brief B → Brief A as the unblock, but both were dead-on-arrival until an\nupstream bug — tracked elsewhere — was fixed; that bug was the true head and wasn't even in the path.)\nBefore you finalize the order: for each \"smallest unblocking move,\" **verify it actually works** —\ncheck the code\u002Fconfig\u002Fbug tracker, don't assume. If a proposed first step depends on something\nunproven, that something is the real head. Say so explicitly.\nThe requester's stated edges are a claim to check, never the evidence for the head, and the check\nis done now, while you author — never left to brief 01's Task. Do it against what you can read in\nthis session: the code, config or tracker when you have them, and always the requester's own words\nand the material you were handed. Test each stated edge there: does the step it puts first rest on\nanything that text does not establish? A run without repo access still has that text, so \"I could\nnot read the repo\" names what you did not look at; it is never the check.\nState the head in one passage, in your own prose, that names it, quotes the line the check rests\non with where it came from, and says what that line showed, in one of two forms:\n\"I verified \u003CX> is the real head by reading \u003Csource>: '\u003Cline>'; it showed \u003Cwhat>\", or\n\"\u003Csource> says '\u003Cline>', which does not establish \u003Cwhat>. \u003CX> is unproven, so it is the real\nhead: brief 01 settles it\". \"That makes 01 the head\", \"established from the stated edges\", and an\n\"unproven\" with no line read and quoted state no check.\n\n## Carry quality obligations into execution\n\nCarry applicable quality attributes and cost-model obligations from the controlling spec\ninto the brief's existing Context, Task, DoD and Verify fields. Name the affected path or\nquality scenario, operating conditions, acceptance target or unresolved decision, behaviour\nto preserve, and evidence owner and method. For performance, retain workload and budget. Cite the source; do not silently invent or relax\nbudgets. If the source omits a consequential constraint, surface the gap before committing\nwork that depends on it. Keep unrelated briefs free of boilerplate quality checklists.\n\nWhere algorithm or data structure choice is consequential, state what needs comparison and\nwhy. An unresolved choice that could invalidate downstream work needs a bounded spike ahead\nof that work, with a decision criterion and explicit dependencies. Verification should cover\napplicable resource budgets and scaling as well as functional equivalence; a faster result\nthat skips required behaviour does not pass. Reuse the spec's workload and measurement\nconditions so that the evidence answers the original requirement. Keep cross-cutting\nobligations traceable to the briefs that implement them and the brief that verifies the\nend-to-end result; passing isolated component checks is insufficient for a system target.\n\n## Brief structure (one file per unit of work)\n\n`docs\u002F\u003Cphase>\u002Fbrief-\u003CNN>-\u003Cslug>.md` (or whatever your repo's convention is — match existing phase docs).\nConceptually every brief carries the same handful of things: **Wave, Depends on, Unblocks, Effort**\n(the scheduling metadata), **Context** (why + the exact facts needed), **Read first** (required\nreading), **Deliverables** (concrete outputs), the **Interface contract** (the seam other briefs rely\non), and a **Definition of Done**.\n\nOlder brief sets narrated all of this in prose — a header line plus paragraphs. That drifts in two\nspecific ways once a plan has more than a few briefs: dependencies written as free-text arrows that\ndon't survive a rename or a re-numbering, and DoD items with no runnable check attached, so \"done\" is\nwhatever the implementer felt like asserting. **The template below is the current standard — use it\nfor every new brief.** It carries the same concepts as data (YAML frontmatter fields, tables) instead\nof only prose, so a script — or the next agent, or you in six months — can rely on the structure\nwithout re-reading the whole plan to extract it.\n\n**Start here: don't hand-type a brief — generate it.** `statusgen newbrief --stream \u003Cs> --title \u003Ct>`\nemits this skeleton with every derivable field already correct: the gate computed from the four risk\nanswers (it refuses a supplied gate, and in non-interactive mode refuses an unanswered risk question\nrather than defaulting it), the wave computed from `--depends`, the inverse `unblocks:` edge written\ninto each named dependency in the same change, and a freshness stamp from a fetch it performs. Hand-\nauthoring stays valid (fix a typo without re-running a tool); the generator just removes the fields an\nauthor can otherwise get wrong. Run `statusgen newbrief --help` for the flags.\n\n### Brief file template (data-first)\n\nSelect the identifier\u002Fschema pair before writing. With negotiated prefixed-label\nsupport, write `brief: \u003Ccell>:\u003Crepo>:\u003Cstream>:b-\u003CNN>` and keep the authored revision\nin `version:`. Dependency refs may use `:b-\u003CNN>@\u003Cversion>`; no suffix binds the selected\nsource snapshot. Use that supporting schema for references\nand dependency fields. The template below shows the currently supported legacy\n`brief-v2` pair; do not combine its schema with an unsupported `b-` identifier.\nThe unprefixed label is deprecated for new authoring after adoption, not invalid for\nlegacy consumers. Existing canonical fullIds and filenames remain unchanged (brief §3.8).\n\n```markdown\n---\nbrief: \u003Ccell>:\u003Crepo>:\u003Cstream>:\u003CNN>   # current brief-v2 fullId; board label is \u003Cstream>\u002F\u003CNN>\ntitle: \u003Cone line>\nwhy: \u003Cprose>                        # REQUIRED for every NEW brief — one to three lines a non-engineer\n                                    # could read and justify the work from. Not the WHAT (that's title\n                                    # + Task), but WHY it matters: what problem, whose pain, what the\n                                    # outcome unlocks. Quality bar: \"all 22 pages are eagerly imported\n                                    # into a single JS chunk, so every user pays a multi-MB first load;\n                                    # splitting cuts the initial payload ~5×.\" (frontend\u002F05 exemplar).\n                                    # Omit only on legacy briefs being backfilled later — statusgen\n                                    # NOTICEs a missing why: (non-fatal this phase; hard error once\n                                    # backfill lands — same pattern as gate-why).\nwave: \u003Cint>\ndepends: []                        # typed IDs only: [\"\u003Cstream>\u002F\u003CNN>\", ...] — NEVER prose arrows\nunblocks: []                       # typed IDs\neffort: S | M | L                   # closed scale — L is the CEILING (rule 12); XL is not a size,\n                                    # it is the signal to split into multiple briefs, each ≤ L.\ngate: model | human                # from the four risk questions below\nrisk: {regulatory: no, customer: no, irreversible: no, sensitive-data: no}\nissues: []                         # GH issue numbers this brief closes\nregression-of: \u003Cissue ref or sha>  # OPTIONAL — fix\u002Fbug briefs only (rule 14): the PRIOR fix this\n                                    # defect re-opens — its issue reference or its commit sha.\n                                    # Omit when no earlier fix exists (a first occurrence); never\n                                    # invent one to fill the field. No lint reads it yet: the key\n                                    # is tolerated as an unrecognised frontmatter key, not validated.\nschema: brief-v2                    # required in a v2 tree;\n                                    # --lint PROBLEMs a tree of v2 briefs missing it. Reserved,\n                                    # OPTIONAL keys parsed under brief-v2 (shape-validated only,\n                                    # gating behaviour deferred to the graph stream): id (uuid,\n                                    # minted once at authoring, never reused), supersedes ([]),\n                                    # version (int, bumped on every Task\u002FVerify edit after first\n                                    # dispatch), gates ([{on, type, reason}]), feathers ([...]).\n                                    # NO status: KEY — the lifecycle cell is DERIVED by statusgen\n                                    # from PR trailers, verifyrun witnesses and App approvals, never\n                                    # hand-asserted in frontmatter or the stream README table.\nauthored: \u003CYYYY-MM-DD> by \u003Cwho\u002Fsession>\nsources: []                        # provenance: scoping doc, finding IDs, intake IDs this derives from.\n                                    # F-\u003Cslug>\u002FI-\u003Cslug> references should be links: \"[F-ws-token-expiry](..\u002Ffindings\u002FYYYY-MM-DD-example-finding.md)\"\ngate-why: \u003Cprose>                  # REQUIRED when gate: human OR any risk answer is yes — what about THIS\n                                   # brief makes it risky and what the human is confirming (a line or two).\n                                   # Omit only when gate: model and all four risk answers are no.\ndecision-trigger: creation | start | spec   # OPTIONAL, gate: human only (absent = start) — WHEN this\n                                    # brief's self-contained decision issue is filed by the\n                                    # `\u003Cdecision-issue filer>` your repo's instructions file names\n                                    # (see rule 4; the same shape as `human:\u003Cname>` — a capability the\n                                    # adopter binds, not a tool this skill ships). A JUDGEMENT recorded\n                                    # per brief, not a global rule: `creation` = the options are already\n                                    # enumerable as authored (file on landing); `start` = the decision is\n                                    # concrete once a worker first picks the brief up — the default,\n                                    # filed by the dispatcher; `spec` = the options only become\n                                    # well-formed after the pickup-time design\u002Fplan step — the executor\n                                    # authors `## Human decision` from that step, and THAT files it.\nexec-tier: any | strong             # OPTIONAL (absent = any) — minimum execution-model tier.\n                                    # DERIVED from four complexity questions (see rule 9); any yes → strong.\nexec-tier-why: \u003Cone line>           # Recommended when exec-tier: strong — which question(s) it answered yes.\ndomain: clear | complicated | complex | chaotic   # OPTIONAL (absent = complicated) — the work's Cynefin\n                                    # domain. Governs which diagnostic to reach for: Ordered\n                                    # (clear\u002Fcomplicated) → Theory of Constraints; Complex →\n                                    # probe-sense-respond \u002F enabling constraints; Chaotic → act first,\n                                    # then sense. `statusgen --lint` PROBLEMs an unrecognized value;\n                                    # `statusgen --cynefin` shows the distribution and lists an\n                                    # absent-domain brief as Disorder. A repo that keeps a longer\n                                    # treatment binds it as `\u003Ccynefin-lens explainer>` in its own\n                                    # instructions file; the four values above are the whole contract.\nparallel-streams: []                # OPTIONAL (absent = one worker per brief — the default, and what\n                                    # nearly every brief uses). Declares that this brief's work\n                                    # decomposes into concurrent SHARDS, each scoped to file globs:\n                                    #   parallel-streams:\n                                    #     - {name: engine, files: [\"statusgen\u002F**\"]}\n                                    #     - {name: docs,   files: [\"docs\u002F\u003Cstream>\u002F**\"]}\n                                    # Presence is a REQUEST, never a permission — `statusgen shardcheck`\n                                    # decides. See \"Intra-brief splits\" below before writing it.\nconsumers: []                       # OPTIONAL — list of consumer sites when this brief changes a shared\n                                    # value. Each entry: \"\u003Cpath>: fixed-here | follow-up \u003Cstream\u002FNN> |\n                                    # out-of-scope (\u003Cwhy>)\". Absent = no shared value changed (the default).\n                                    # An AUTHORING PR declares future consumers, it does not edit them:\n                                    # route a path this brief's implementation will LATER touch as\n                                    # `follow-up \u003Cstream>\u002F\u003CNN>` at THIS brief, not `fixed-here`.\n                                    # See rule 6. Authoring convention — no lint enforces it yet\n                                    # (tracked as a planned follow-up).\n---\n\n# Brief \u003CNN> — \u003Ctitle>\n\n## Context\nfiles: \u003Cexact paths the implementer touches>\nout-of-repo files: \u003Cexact paths outside the repo (e.g. ~\u002F.claude\u002Fskills\u002F...), if any —\nsee rule 7; omit the line entirely when none>\nfacts: \u003Cthe 3-5 project facts needed — key: value, no narrative. The implementer\nmust never need to explore the repo.>\nlayering: \u003CREQUIRED for a new component\u002Fservice\u002Ftool or a substantial boundary change\n(domain extraction\u002Fdissolution, added adapter\u002Fentrypoint, logic crossing a process or\ntrust boundary); omit otherwise. One or two lines: simplest justified structure +\nmeaningful rules vs external effects (or why no extraction is warranted) + current\nreason + Task\u002FVerify references. Defaults (domain-core \u002F flat tool), contract placement,\ninterface\u002Fpurity rules, boundary checks and existing-brief amendment policy: brief-v1 §4.1.>\ndesign-fit:                  # REQUIRED on every NEW brief; its two rules: brief-v1 §4.1\n  owner: \u003Cthe one module that owns the meaning this brief touches, or n\u002Fa>\n  contract: \u003CS-\u003Cslug> row id of the semantic-owner index, or none — \u003Cwhy>>\n  retires: [\u003Cmechanism\u002Frefusal\u002Fflag\u002Ftest this brief removes>, ...]   # [] is an answer\n  weight: \u003Csigned delta per dimension: verbs, flags, refusals, rule-text lines>\n  why-add: \u003Crequired when any delta > 0: why it cannot live in the owner, and what\n           removal was considered instead; n\u002Fa otherwise>\noracle: docs\u002Finvestigations\u002F\u003Cfile>  # ONLY on a design brief a brittle `redesign` raised: its refactor\n                                    # oracle, whose five §5 rows Verify copies. Such a brief without it is not authored\n\n## Human decision\n\u003C!-- gate: human only — omit the section entirely otherwise. Lifted VERBATIM into the\n     brief's decision issue by the `\u003Cdecision-issue filer>` your repo's instructions file\n     names, so write it SELF-CONTAINED: the reader sees ONLY this text, never the brief —\n     no links, no URLs (a conforming filer refuses both), no repo paths, no references to\n     other briefs or docs. REQUIRED at authoring time when decision-trigger is creation or\n     start; when decision-trigger: spec, the EXECUTOR authors it at pickup from the\n     design\u002Fplan step and commits it into this file. -->\n\u003C3-10 lines: what is being decided and why it needs a human — enough to decide from\nthis text alone>\n\nOptions:\n1. **\u003Coption>** — \u003Cwhat it means; consequences; what happens next if chosen>\n2. **\u003Coption>** — \u003Csame>\n\nDefault if no answer: \u003Coption N after \u003Cdate\u002Fwindow>, or \"none — blocks until answered\">\n\n## Ground rules\n- NEVER git push \u002F trigger workflows \u002F run mutating kubectl. Leave commits per\n  the task instructions only.\n- Stop at `implemented` — you do not set verified\u002Fdone.\n- If anything is unclear or contradicts repo state: report NEEDS_CONTEXT, don't guess.\n\n## Task\n\u003Cwhat to build\u002Fchange — explicit steps>\n\n## Verify (executable — no prose-only DoD items)\n| # | Command | Expect |\n|---|---------|--------|\n| 1 | \u003Cliteral command> | exit 0; output contains \"\u003Cliteral>\" |\n\n## Evidence\n\u003C!-- appended at implementation time: one row per Verify item —\n     (command, exit code, output line(s) or hash, date, runner).\n     \"verified\" status in the stream README requires this section filled\n     by someone who did NOT implement. -->\n\n## Review\nGate: \u003Cmodel|human> (from frontmatter). Reviewer records verdict + date in the\nstream README table. Human gate is MANDATORY when any risk answer is yes.\n```\n\nFill every frontmatter field — an empty `sources: []` or a missing `risk:` answer is not a shortcut,\nit's a gap the reviewer will bounce back. `gate` is derived, not chosen: if any of the four risk\nquestions is `yes`, `gate` must be `human`; only when all four are `no` may `gate` be `model`.\n\n### Rules behind the template\n\n1. **Load-bearing facts live in islands, not prose.** Frontmatter, tables, and Verify rows are the\n   record; prose explains and motivates but never solely carries a fact a script or a reviewer needs.\n2. **References are typed IDs** (`stream\u002FNN`, `F-\u003Cslug>`, `I-\u003Cslug>`, legacy `F-NN`, `I-NN`, …) — no\n   fuzzy names (\"the pricing brief\"), no arrow notation (`←12a`). A typed ID survives renumbering\n   and greps cleanly. **Register IDs are letter-prefixed slugs:** `F-\u003Cslug>` for\n   findings, `I-\u003Cslug>` for intake — the slug is 10–20 chars, `[a-z0-9-]`, derived from the title\n   (example: `F-ws-token-expiry`, `I-model-mix-tiers`). Older entries use numeric IDs (`F-01`, `I-01`)\n   — those are frozen legacy and remain valid; new entries must use the slug form. **Going\n   forward, register references SHOULD be clickable links pointing at the per-entry\n   file: `[F-ws-token-expiry](..\u002Ffindings\u002FYYYY-MM-DD-example-finding.md)`. The typed ID\n   remains the primary key (link text); the link is for readers. The mapping is each entry file's\n   frontmatter `id:` — never guess slugs. Bare (unlinked) IDs stay legal but are not the preferred\n   form for new briefs.\n3. **Verify rows must be runnable by someone who didn't do the work.** A DoD\u002FVerify item with no\n   literal command and no expected exit\u002Foutput is not a DoD item — it's a hope.\n   Prefer POSIX commands (`grep -F`, forward-slash paths) for portable checks. For a genuinely\n   native Windows command, add a `Shell` column and explicitly mark the row `cmd` or `pwsh`\n   when authoring it. Never emit `findstr \u002Fc:\"…\"` with backslash paths under implicit `sh`:\n   bash reinterprets those arguments. Shell selection belongs in the reviewed Verify table;\n   an Evidence-only edit must not change the command or its shell to make a failing row pass.\n   **Say where each row runs.** When a row needs anything beyond the checkout, add an `Env`\n   column (`| # | Class | Env | Command | Expect |`): `tree`, `host[:\u003Cname>]`,\n   `service:\u003Cname>` for a named NON-PRODUCTION instance (a database, an API, a staging\n   deployment, described in `## Context`), or `cluster:\u003Cname>`. A row whose Env is not `tree`\n   is never `check:ci`, since no CI job reaches it; class it `check`. A row needing a project\n   secret or a live service is witnessed by a non-implementer where that environment exists,\n   never by the implementer's own report. The grammar and its `statusgen --lint` rules are\n   rule 45 in the brief-rules reference.\n4. **`gate: human` is mandatory when any risk answer is `yes`.** Record all four risk answers\n   (regulatory, customer, irreversible, sensitive-data), not just the resulting gate — the answers are\n   what a reviewer audits, the gate is just their conclusion. **A risk-gated brief (`gate: human` OR\n   any risk answer `yes`) MUST carry a `gate-why`** — a line or two naming what about *this* brief\n   trips the wire and what the human is actually confirming at the sign-off. A risk boolean with no\n   rationale is an unfalsifiable assertion; the verify-gate card surfaces the `gate-why` verbatim.\n   (statusgen NOTICEs a missing `gate-why` today; it becomes a hard lint error once the backfill lands.)\n   **A gate:human brief also gets a filed DECISION ISSUE.** The human gate is only real when there is\n   a concrete, self-contained tracker issue for the human to act on. Two halves, and only the first\n   ships with this methodology: `statusgen --decision-issues` emits the newly-eligible set as JSON —\n   title, `needs-decision` label, composed body, and a hidden `\u003C!-- needs-decision: \u003Cstream>\u002F\u003CNN> -->`\n   marker that dedupes, so a brief that ever had one never gets a second — and the\n   `\u003Cdecision-issue filer>` your repo's instructions file names is what actually files it. (Same shape\n   as `human:\u003Cname>`: a capability the adopting repo binds, because filing depends on the tracker.)\n   Record the resulting number back in the brief's `decision-issue:` frontmatter field; `statusgen\n   --lint` NOTICEs a gate:human brief in flight without one. WHEN it files is a per-brief judgement\n   you record as `decision-trigger:` (see the template): ask *\"could the reader pick an option from\n   what I can write down right now?\"* — yes at authoring → `creation` (author the `## Human decision`\n   section now and file as the brief lands); yes once a worker frames the work → `start` (the default;\n   the dispatcher files it, from your authored `## Human decision` section, or by deriving a fallback\n   from `gate-why`\u002F`why` when the section is absent); only after real design exists → `spec` (the\n   executor authors the section at pickup, which is what ripens the filing). The issue body must let\n   the human decide WITHOUT opening the repo — inline everything, since a conforming filer refuses\n   links and URLs in the composed body.\n5. **Provenance is required.** A brief with an empty `sources:` — no scoping doc, no finding ID, no\n   intake ID it derives from — is untraceable: no one can tell why the work exists or whether it's\n   still needed.\n6. **A brief that changes a SHARED VALUE must enumerate its consumers, and verify the FLOW — not just\n   the site.** (Cost that forced this rule: an API response field was renamed and verified green at its\n   own site — \"the endpoint returns the new field name\" — while a frontend that still read the *old*\n   field name silently broke the whole end-to-end flow; no dashboard row ever rendered. The change was\n   correct locally and broken across components.) A shared value is anything another component reads:\n   a party\u002Fidentity, an env-var NAME, a ConfigMap\u002FSecret key, a field's MEANING, a wire\u002FJSON format,\n   or a default. The trigger is cross-component: when the brief's Context `files:` references multiple\n   top-level directories (k8s\u002F, api\u002F, frontend\u002F, services\u002F, docs\u002F, docs-site\u002F, scripts\u002F, .github\u002F,\n   tools\u002F) or its Task touches a shared-surface keyword (party, identity, env-var, ConfigMap, Secret,\n   wire format, JSON format, a default).\n   - **Enumerate consumers as a structured frontmatter field**:\n     `consumers:` — an optional YAML list in the brief-v1 frontmatter (absent = the brief asserts it\n     changes no shared value). Each entry is `\u003Cpath-or-component>: \u003Crouting>` where routing is one of\n     `fixed-here` \u002F `follow-up \u003Cstream>\u002F\u003CNN>` \u002F `out-of-scope (\u003Cwhy>)`. An unlisted consumer is how\n     you strand an assumption. Project wrappers (this repo's `.claude\u002Fskills\u002Fauthor-brief\u002FSKILL.md`)\n     carry the full format spec. **The lint enforces this field** — `statusgen --consumers` is the\n     diff-aware gate (it exits non-zero on a routing claim the diff disproves), and `statusgen\n     --lint` alone already makes one class **fatal**: a `follow-up \u003Cstream>\u002F\u003CNN>` naming a brief that\n     exists in no stream README is a hard PROBLEM. The rest of the field's checks are advisory\n     NOTICEs. See the generated **Enforcement status of these rules** table below for the exact\n     per-rule status; do not restate it here (it drifts — spec §4 B9).\n   - **Never reclassify an entry to make it fit the grammar.** The routing token is a commitment\n     record — `follow-up` says work is still owed, `out-of-scope` says it isn't — so swapping one\n     for the other to satisfy a regex destroys the fact the field exists to hold. A `fixed-here`\n     wanting a `(\u003Cwhy>)` parenthetical is already legal — the grammar is tolerant of trailing detail\n     (`fixed-here (bumps the pin in the same commit)`), so write it. The remaining awkward case is a\n     `follow-up` whose target is an intake entry or an unauthored brief with no `\u003CNN>` to cite: do\n     **not** leave a bare `follow-up` on the belief nothing enforces it — a `follow-up` with no\n     target is a NOTICE, and inventing a `\u003Cstream>\u002F\u003CNN>` that does not exist is a **fatal** PROBLEM.\n     Route it truthfully as `out-of-scope (\u003Cwhy — deferred to an unauthored brief \u002F intake entry>)`,\n     which records the deferral without a false or missing target. Same reason existing briefs are\n     not backfilled: the record's truth outranks the grammar, and the grammar has a truthful form.\n   - **An authoring PR routes its OWN future deliverables to the deferred disposition, never\n     `fixed-here`.** A brief-authoring change *declares* where a change will land; it does not land\n     it — the authoring diff adds the brief and touches none of the paths the brief's own\n     implementation will later edit. `fixed-here` is a claim about THIS diff, legitimate only when\n     THIS diff actually edits the named path; asserting it on an authoring-only diff is a claim the\n     routing gate correctly disproves — \"no path it resolves to appears in the diff\" — reddening the\n     PR on its first push. So route each path this brief's own implementation will later touch to the\n     **deferred disposition**: `follow-up \u003Cstream>\u002F\u003CNN>` targeting THIS brief. The implementation\n     change that lands the path then replaces that routing with `fixed-here`, in the same change.\n     A self-targeting `follow-up` corroborates by construction — the brief's own file contains its\n     own ID, so the gate's back-reference check is satisfied the moment the brief exists. Every entry\n     still ends in exactly one of the three routing tokens; an entry with none claims nothing and can\n     corroborate nothing. Worked example — the same consumer written wrongly then rightly on an\n     authoring PR that only files the brief:\n     - wrong (asserts a fix this diff did not make; the gate disproves it): `- \"src\u002Fpipeline\u002Frouter.go: fixed-here\"`\n     - right (deferred to this brief until its implementation lands): `- \"src\u002Fpipeline\u002Frouter.go: follow-up \u003Cstream>\u002F\u003CNN> (this brief; flips to fixed-here when the implementation edits the path)\"`\n     Run your project's consumers check — the diff-aware `statusgen --consumers` gate — over the\n     branch before pushing an authoring change, so the deferred routing is confirmed locally rather\n     than by a red check in CI.\n   - **The consumers Verify row works after the merge too**: the bare\n     `statusgen --consumers --brief \u003Cstream>\u002F\u003CNN>` row re-runs on merged main with no hand-pinned\n     commits. It pins itself to the delivery and that commit's first parent, read locally: the\n     committed Evidence's `Verification-Table:` marker if there is one, otherwise the one\n     first-parent commit whose message (or a commit it merged) carries `Brief: \u003Cstream>\u002F\u003CNN>`. So\n     the implementation PR's commits carry that trailer, and a later PR for the same brief should\n     not repeat it: zero or several candidates is exit 2 (COULD-NOT-CHECK), and the remedy is an\n     explicit `--delivery \u003Cfull-sha> --base \u003Cits first parent>`.\n   - **Flow Verify row**: a shared-value brief's Verify table must carry at least one row exercising\n     the cross-component flow end-to-end (e.g. *frontend submits form → API persists → dashboard row\n     appears*), not just the changed site. Site-green does not imply flow-green. This one stays a\n     judgement call even after the planned check ships — prose Verify rows cannot be perfectly\n     classified, so the check is an advisory NOTICE, never a hard gate.\n   - **Cross-repo pairing**: when a change crosses a repo boundary in addition to a component\n     boundary, it carries BOTH a `consumers:` list (this rule) AND a sibling-repo SHA — note\n     the pair in both places.\n   This applies to **issue-driven work too** — a change to a shared value that skips a brief skips\n   this analysis entirely, which is the highest-risk path in the system. Promote such a change to a\n   brief, or at minimum run the consumer grep and record it.\n7. **Out-of-repo files must be declared — the declaration is the claim.** A brief whose scope\n   touches files outside the repo (user-level skills\u002Fsettings, e.g. `~\u002F.claude\u002F**`) MUST list the\n   exact paths in Context under `out-of-repo files:`. These files have no worktree isolation and no\n   branch-as-claim, and an edit takes effect in every running session the moment it's written —\n   review is post-hoc. So: the dispatcher **serializes** on the declaration (at most ONE such brief\n   in flight at a time, across all streams), and the implementer stages the edit as a diff in the\n   PR, applying it to the live file only as the LAST step before `implemented` — never\n   incrementally. Where the out-of-repo surface is itself a git repo (e.g. the `~\u002F.claude` stopgap\n   repo), commit the applied edit there too. A brief that touches this surface without the\n   declaration is scope drift — bounce it.\n8. **Freshness check — before authoring a fix\u002Fchange brief, verify the site isn't already\n   fixed.** Read the target site on fresh `origin\u002Fmain` (`git fetch origin && git log\n   origin\u002Fmain --oneline -- \u003Cpath>` + read the code) and record the check in the brief's\n   `sources:` as `freshness-checked \u003CYYYY-MM-DD> @ \u003Cshort-sha>`. If the deliverable is already\n   satisfied, the brief is NOT authored — resolve or re-scope the finding\u002Fissue instead. That\n   is the only reason to withhold it: a read stated to you as done counts as done (stamp its\n   sha), and a result you cannot fully see is no reason to hold — author the brief, with a first\n   Task step that re-checks. The output is the brief or a \"do not author — already fixed\"\n   determination, never a held draft or \"not authored yet\". The check is light; the waste it prevents (re-writing a fix that already landed, claiming a vuln\n   closed that was never open) is heavy.\n9. **exec-tier: complexity signals a minimum execution-model tier.** `exec-tier` is DERIVED,\n   not chosen (mirror of the gate\u002Frisk rule). The author answers four complexity questions;\n   any yes → `exec-tier: strong` (absent\u002Fany-no → `any`, the default):\n   (a) Does the Task require design decisions the facts do not fully pre-specify?\n   (b) Does correctness depend on cross-component\u002Fcross-artifact reasoning (shared values,\n       end-to-end flows, sweeping a pattern across sites)?\n   (c) Is it code where a subtle implementation error survives the brief's own tests (auth,\n       funds, concurrency, safety plumbing)?\n   (d) Is this a design brief raised by an error-class trigger? (A recurring defect class, not one symptom, has accrued enough counted instances or merged fixes to be owed a design.)\n   `strong` SHOULD carry a one-line `exec-tier-why` naming the question(s) (lint: PROBLEM if unknown, NOTICE if missing).\n   **Honest limitation:** statusgen never verifies which model actually ran — pickup-side\n   compliance is honor-system self-report; the enforced surface is desk dispatch + the review gate.\n10. **Defense in depth is the DEFAULT design posture.** When a brief's Task has design\n    alternatives, prefer the LAYERED one: no single control standing between a fault and the\n    damage. **Mandatory for core-system briefs** — each project layer defines what \"core system\"\n    means for its own repo (typically: anything touching money, auth\u002Fidentity, or an irreversible\n    external write); recommended everywhere else. Three authoring obligations:\n    - (a) **Single-point-of-failure note in `## Context`** — one or two lines answering \"what is the\n      ONE control this design depends on?\". If the honest answer is a single thing, the design needs\n      another layer BEFORE the brief is authored, not after a reviewer bounces it.\n    - (b) **At least two INDEPENDENT layers in the Task**, where feasible. **The independence test:\n      two layers are independent only if they fail for DIFFERENT reasons in DIFFERENT components.**\n      The same check pasted into N call sites is one layer — it fails once and everywhere. A typical\n      independent set: an enforcement point at the trust boundary (the system itself rejects the\n      operation) + validation in the service that calls it + an out-of-band monitor\u002Freconciler that\n      notices when both were bypassed.\n    - (c) **Verify rows exercising each layer INDEPENDENTLY, including at least one NEGATIVE-PATH\n      row** — break or bypass the upper layer and prove the lower one still catches the fault. A\n      Verify table that only walks the happy path through every layer at once has verified exactly\n      one layer: the outermost.\n    **Boundary — this is NOT \"add more assertions\".** Layers are DESIGNED controls at DISTINCT trust\n    boundaries, not one check duplicated for reassurance. Where a project's constraints say invariant\n    checks are closed-form math rather than runtime assertions, that constraint stands; the\n    independence test in (b) is what distinguishes a real layer from assert-spam. Adding a \"layer\"\n    that fails for the same reason as the layer it backs up costs review time and buys nothing.\n11. **A brief whose deliverable makes checkable factual claims needs at least one DEREFERENCING\n    Verify row, not only presence\u002Fformatting checks.** Two rows can both go green without proving\n    the same thing:\n    - **Presence\u002Fformatting** — `grep -c`, `wc -w`, \"section exists\" — proves the deliverable is\n      well-formed. It CANNOT fail on a wrong-but-well-formed document: a confident falsehood sitting\n      in the right section, at the right length, passes exactly like the truth.\n    - **Dereferencing** — fetch a link and check what it actually serves; run a documented command\n      and compare its real output\u002Fexit code to the specific claim; check a documented ID or property\n      against the live system it names. This CAN fail on a wrong-but-well-formed document.\n\n    Rule 3 above (\"Verify rows must be runnable by someone who didn't do the work\") still stands —\n    this rule doesn't relax it. It says that when the deliverable itself carries checkable facts (a\n    setup\u002Fconfig guide, a market or competitive analysis, a spec — anything asserting \"X is true\"\n    that a reader could act on), a presence gate can't be the *whole* mechanical floor: at least one\n    row has to be capable of catching a wrong claim, not just a missing section. Skip this only for\n    genuinely presence-only deliverables with no factual claim to dereference (a docs reformat, a\n    template scaffold) — manufacturing a dereferencing row against a deliverable with nothing to\n    check is its own kind of checkmark-DoD.\n\n    It is a DIFFERENT AXIS from the row-runner rules and the unfailable-row lint: those read the\n    command TEXT and are enforced by `statusgen --lint`; this one needs to know what the deliverable\n    claims, so no lint enforces it. A row can pass every one of them and still measure only presence.\n\n    The triggering evidence (an all-presence table that passed 8\u002F8 on a factually wrong guide) and\n    why a link-resolution lint would cover only link-shaped dereferencing are in the brief-rules\n    reference, rule 43. Don't wait for such a lint to exist before writing the row this rule asks for.\n12. **Sizing limits — L is the hard ceiling; prefer M for strong-tier \u002F risk-gated work.** The\n    `effort:` scale is closed at S | M | L, and the top of it is a rule, not a convention: **no\n    brief may be larger than L. A unit of work that would be XL (or bigger) MUST be split into\n    multiple briefs, each ≤ L**, sequenced with `depends:`\u002F`unblocks:` (or sharded via\n    `parallel-streams:` when the intra-brief split rules below hold). XL is only ever a\n    signal-to-split, never an emitted size — `statusgen --lint` already PROBLEMs any effort\n    outside S\u002FM\u002FL.\n\n    **Preference, explicitly NOT a rule:** when a brief is `exec-tier: strong` OR risk-gated\n    (`gate: human` \u002F any risk answer `yes`), **prefer M as the ceiling** — lean toward splitting\n    an L into Ms. An L remains permitted for strong\u002Fhigh-risk work when there is a real reason to\n    keep the unit whole; the default lean is M-or-smaller. Rationale: strong-tier work\n    (cross-artifact correctness) and high-risk work are exactly where a large brief most degrades\n    review quality, so smaller, more-reviewable units are preferred there — the review gate is\n    the control the risk answers lean on, and it only bites on a diff a reviewer can actually\n    hold.\n\n13. **A Verify row MAY declare an OBLIGATION alongside its class — the prose MUSTs of rules 6, 10 and\n    11 carried as typed data.** The optional `Class` cell (rule 45 in the brief-rules reference) answers\n    WHO EXECUTES a row (`check:ci` \u002F `check` \u002F `gate:model` \u002F `gate:human`). A second, orthogonal axis\n    answers WHAT OBLIGATION the row discharges, written as `+`-prefixed tokens appended to the same\n    cell — a **compound cell**, `\u003Cexecution> +\u003Cobligation> …`, so the table's column set is unchanged\n    and a legacy column-less table is untouched. Four obligation values, each an example row:\n\n    | Obligation | What the row must do | Example `Class` cell |\n    |---|---|---|\n    | `+mutation` | break the thing the change guards and prove the guard reddens (rule 10c, spec D1) | `check:ci +mutation` |\n    | `+flow` | exercise the cross-component path end to end, not just the changed site (rule 6) | `check +flow` |\n    | `+dereference` | resolve a claim rather than count its presence (rule 11) | `gate:model +dereference` |\n    | `+neighbour` | exercise a sibling site the change did not touch | `check +neighbour` |\n\n    `statusgen --lint` derives which obligation a brief OWES from the shape of its change (the branch\n    diff, its declared paths, its task prose) and reports one owed-but-absent — it checks the PRESENCE\n    of the typed row, never its ADEQUACY, which stays the reviewer's call (spec §3 D7). An unknown\n    `+token` is a hard PROBLEM, exactly as an unknown class is. **Enforcement status:** the derivation\n    lands advisory (a NOTICE); `+mutation`'s promotion to a hard gate is a follow-up. This status line\n    is deliberately minimal — the authoritative per-obligation enforcement state is generated, not\n    hand-copied here (a hand-written status becomes the next stale second copy).\n\n14. **A fix\u002Fbug brief closes the defect CLASS, not the one instance.** A defect repaired at one\n    site comes back at another when the fix closed the instance and left the class open: a second\n    caller reaches the same hazardous primitive by a different path, a test stub hides it, and the\n    regression reads as a new bug. So a brief whose Task fixes a defect carries two things beyond\n    the ordinary template:\n    - **`regression-of:`** in the frontmatter, when an earlier fix for the same defect exists —\n      that fix's issue reference or commit sha. It records that the class has escaped once\n      already, and it points the reviewer at the guard that failed to hold. Omit it for a first\n      occurrence; never invent a value to fill it. Run the freshness check (rule 8) over the\n      earlier fix's site as well as the new one.\n    - **A MANDATORY class-guard Verify row.** Name the class in `## Context` `facts:` — the shape\n      every instance shares, not the one line that failed — and add a row whose check covers EVERY\n      site the defect can recur at, not only the reported one. The model is an allow-list\n      structural test: enumerate every caller of the hazardous primitive (`exampleRawToken()`) and\n      fail on any caller outside a short committed allow-list (`exampleSafeToken()`). A lint rule, a\n      type that makes the hazardous call unrepresentable, or a single choke point serves equally. A\n      test of the reported instance may sit beside it; it never replaces it.\n    - **The class-guard row is FAIL-FIRST.** Before the fix lands, the implementer plants a\n      deliberate SECOND instance of the defect at a site the fix does not touch and shows the row\n      RED on it; the Evidence records that red run, naming the planted site, beside the green one.\n      A guard shown red only against the reported instance proves it sees that instance, which the\n      instance test already did. Where the guard's matcher could silently stop matching, keep the\n      plant as a committed positive-control fixture the guard must flag, so the red stays\n      re-runnable after merge. Tag the row `+mutation` in its `Class` cell (rule 13) — breaking\n      the guarded property and proving the guard reddens is exactly that obligation.\n\n    Worked example (neutral names) — the frontmatter line and the two class-guard rows:\n\n    ```markdown\n    regression-of: \u003Cthe earlier fix's issue reference or commit sha>\n\n    | # | Command | Expect | Class |\n    |---|---------|--------|-------|\n    | 2 | `go test .\u002Fexamplepkg\u002F -run 'TestExampleRawTokenAllowList$' -count=1` | exit 0 — every caller of `exampleRawToken()` in the tree is on the allow-list | check:ci +mutation |\n    | 3 | `go test .\u002Fexamplepkg\u002F -run 'TestExampleRawTokenAllowListFlagsPlant$' -count=1` | exit 0 — the guard, run over `testdata\u002Fexample-plant\u002F` (one planted caller outside the allow-list), reports that caller | check:ci +mutation |\n    ```\n\n    Row 3 is the positive control that keeps the fail-first re-runnable; the pre-fix red run with a\n    planted second caller in the real tree goes in the Evidence. **Enforcement status:** no lint\n    enforces this rule yet — `regression-of:` is tolerated as an unrecognised frontmatter key rather\n    than validated, and whether a class-guard row is present and truly covers the class is the\n    reviewer's call. A defect with no mechanically checkable shape (a one-off logic error nothing\n    else can repeat) states that in `## Context`, with the reason, in place of the row; the reviewer\n    weighs it, and it is not available for a defect that has already reached a second site.\n\nKeep a brief self-contained: if executing it requires knowledge from another brief, either link it\nunder \"Read first\" \u002F `facts:` or state the dependency in `depends:` — never assume the reader has the\nwhole plan in context.\n\n## Enforcement status of these rules (generated — do not hand-edit)\n\nRegenerate with `statusgen enforcement-status` piped through the skillslint sync\n(`go run .\u002Ftools\u002Fskillslint --sync`); the byte-diff gate in the skillslint CI job fails\nif this block drifts from the lint's rule registry. Hand-edit it and the gate reddens.\n\nThis table reports what `statusgen --lint` enforces, **not** what the methodology\nrequires. A rule enforced only by a tool outside the lint — a CI workflow, a desk-side\nguard — reads as `not enforced` here unless it is registered. The three statuses are\nexact: **fatal** is a `--lint` PROBLEM that makes the run exit non-zero; **advisory** is\na NOTICE that is printed but never gates; **not enforced** means no lint rule checks it —\nit is an authoring convention only.\n\n| rule | what it checks | status |\n| --- | --- | --- |\n| `bre-alternation` | a pipe in a basic-regex grep pattern (no `-E`\u002F`-P`) is an ordinary character, so the pattern matches the Verify row itself — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `cmd-marker-ambiguous` | a Command cell carrying more than one `cmd:`-marked code span, so which command the row names is ambiguous | advisory |\n| `cmd-marker-not-honoured` | a `cmd:` span verifyrun ignores because the rendered cell may not show it as code (backslash-escaped backticks; raw HTML, a link, an image, a dollar in any spelling or a character reference in the cell's prose; a span fused to the text before it; or a marker not set apart by whitespace), so the row runs its first span | advisory |\n| `cmd-marker-overrides-command` | a `cmd:` marker that replaces a first code span which reads as a command itself (multi-word), so the row runs something other than the span a reader sees first | advisory |\n| `cmd-marker-vacuous` | a `cmd:`-marked command that cannot fail (`true`, `:`, `exit 0`, a lone `echo`), so the row passes whatever the tree holds | advisory |\n| `consumers-flow-verify-row` | that a shared-value brief's Verify table carries at least one row exercising the cross-component flow end-to-end — a judgement call no lint decides | not enforced |\n| `consumers-followup-missing-brief` | a `consumers: follow-up \u003Cstream>\u002F\u003CNN>` whose target is not a brief in any stream README — the routing claim is false | fatal |\n| `consumers-followup-no-target` | a `follow-up` routing that names no `\u003Cstream>\u002F\u003CNN>` target — a deferral with no holder | advisory |\n| `consumers-followup-one-way` | a `follow-up` target that exists but never references the deferring brief back — the coverage claim is one-way | advisory |\n| `consumers-missing-list` | a brief whose prose reads as changing a shared surface but enumerates no `consumers:` (a heuristic prompt, never a verdict) | advisory |\n| `consumers-no-verify-row` | a brief carrying `consumers:` but no Verify row that runs `statusgen --consumers` to corroborate the routing | advisory |\n| `consumers-out-of-scope-no-reason` | an `out-of-scope` routing with no substantive reason for the reviewer who must weigh the exclusion | advisory |\n| `consumers-prose` | a `consumers:` written as a prose paragraph rather than a routed list, so nothing can corroborate it | advisory |\n| `consumers-unrouted` | a `consumers:` entry that names no routing token (`fixed-here` \u002F `follow-up` \u002F `out-of-scope`) | advisory |\n| `eligibility-could-not-check` | a gates: edge (or unsatisfied depends:) the eligibility evaluator could not resolve offline — an unregistered\u002Funpublished alias, an absent sibling checkout, or a forge-backed target — which HOLDS the brief out of Next-up | advisory |\n| `ere-literal-pipe` | a `\\|` inside a `grep -E` pattern is a literal pipe, not alternation, so the row matches almost nothing and passes blind — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `fail-first-closure` | a risk-bearing Verify row of a closure this branch makes (verified\u002Fdone, and not so at the merge-base) has no current fail-first witness, is unproven or non-discriminating, was run for an earlier command, or names a base that is not the merge-base or an ancestor of it that already holds the brief at its current path (only the merge-base itself when the brief was created or renamed on the branch) (docs\u002Fverify-row-strength.md) | fatal |\n| `gnu-only` | a GNU-only shell construct that fails on the BSD\u002FmacOS userland a reviewer may run the row on — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `gorun-exit` | a `go run` in the Command cell flattens the program's exit code, so a non-zero result reads as success — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `gotest-run-vacuous` | a `go test -run` selector with no `--- PASS` assertion in the same command, so the row passes whether or not the named test exists, is built, or was ever renamed away (open briefs only; a closed brief's rows are summarised, not individually flagged) | advisory |\n| `grep-pipe-zero-count` | a grep feeding a later pipeline stage under a zero-count Expect fails on its own success path under pipefail (grep exits 1 on no match), and passes on exit status alone when a match is present | advisory |\n| `grep-zero-count` | a `grep -c` whose pass bar is satisfied by a zero count measures nothing — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `moving-ref` | a diff base pinned to a moving ref (a branch name, not a SHA) makes the row's result drift under it — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `no-output-assertion` | an Expect of exactly `exit 0` on a command whose output nobody asserts on (use `exit 0; output contains \"\u003Cliteral>\"` or `exit 0; \u003CN> lines`) | advisory |\n| `pattern-containment-set` | a containment list that is not exactly the five containment values, on a node that is not an effect-free human decision, with a pre-selected default, or a signal pattern without exactly one containment node — containment could become an automated act | fatal |\n| `pattern-culprit-record` | a culprit record whose fields are not the fixed six, or whose default or withheld class is not judgement, or whose withheld file is not null — a withheld or unresolved culprit could route to the mechanical exit | fatal |\n| `pattern-decision-options` | `options`, `default` or `containment` on a node that is not a decision, a default outside the node's options, or a duplicate option — a non-decision node would carry a person's choice | fatal |\n| `pattern-effect-exceeds-role` | a pattern node whose declared effect kind is not permitted for its role, per the role-to-effect-kind table in spec\u002Fworkflow-pattern-v1.md §7 — a generated instance would carry a permission its role does not hold | fatal |\n| `pattern-effect-target-not-owned` | a non-effect-kind pattern node declares an effect whose target is not among its own outputs — a node claiming a consequence outside its own declared output boundary | fatal |\n| `pattern-join-not-check` | a pattern's `join` names a node whose kind is not `check` — the pattern has no independent integration check | fatal |\n| `pattern-judgement-no-default` | a signal class whose judgement exit is not a human decision with options and a default among them — the person closing the decision gets no recommended next step | fatal |\n| `pattern-remedy-set-open` | a signal class listing a remedy outside the closed remedy set, a mechanical exit with no remedy or not kind effect, or no `.github\u002F` refused-path while any node declares effects — an automated fix could touch gate configuration | fatal |\n| `pattern-replay-not-mandatory` | a signal pattern with no node producing the reserved `replay-fixture` artifact, or a producer that owes it under no mandatory command claim — a fix could land with no reproduction of the failure | fatal |\n| `pattern-review-same-role` | a pattern node whose evidence includes a review claim but shares its role with the node that produced its input — the implementer\u003C->reviewer separation is not machine-checked | fatal |\n| `pattern-risk-input-missing-verdict` | a pattern's `risk-input` omits one of the four risk-class verdicts (low\u002Fstandard\u002Felevated\u002Fhuman) — an instance of that risk class has no declared mandatory gates | fatal |\n| `pattern-signal-admission-open` | a signal class admitting a trigger outside schedule\u002Fpush, push from a ref other than refs\u002Fheads\u002Fmain, or no admission at all while any node in the pattern declares effects — a run of unchecked provenance could reach an automated effect | fatal |\n| `pattern-signal-input` | a signal pattern whose signal input is malformed or off an entry node, whose input classes differ from the declared signal classes, that has a node unreachable from the signal input, or whose class names a missing exit or an incomplete gate-red record — a class could escape the rules that range over the input | fatal |\n| `pipeline-exit-sunk` | a shell pipeline whose real exit status is sunk by a later stage, so the row cannot fail — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `prose-led-command` | a Command cell whose lift — the `cmd:`-marked command, else the first code span — is not a command: a prose mention (a file, a path, an `owner\u002Frepo`, a code identifier, a word ahead of the real command), a shell comment, or a shell no-op (`:`, a cmd.exe `REM` \u002F `::` remark); verifyrun and the desk executors record the row could-not-run. Fix a mention by marking the real command with a `cmd:` code span; fix a comment or no-op, marked or not, by replacing it with the command that performs the check | advisory |\n| `rE2-literal-pipe` | a `\\|` inside a `go test -run`\u002F`-bench` selector is a literal pipe in RE2, not alternation — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `runtime-resource-claim-without-principal` | a runtime resource claim (or a release of a claimed resource) not authorized by a human or organization principal, or ownership no succeeded claim backs | fatal |\n| `runtime-resource-credential-not-rotated` | a runtime resource claim with no recorded rotation of the agent credential, or a release with no recorded revocation | fatal |\n| `runtime-resource-credential-outlives-ttl` | a runtime resource whose agent credential lifetime exceeds the resource's ttl | fatal |\n| `runtime-resource-expired` | an unclaimed runtime resource past its ttl, or a claimed one past its retention, still recorded as live | fatal |\n| `runtime-resource-receipt-missing` | a runtime resource effect recorded succeeded with no adapter receipt | fatal |\n| `runtime-resource-schema-violation` | a runtime resource record that breaks runtime-resource-v1's shape, repeats an effect id, or carries a ttl, lifetime or timestamp the rules cannot evaluate | fatal |\n| `runtime-resource-state-unsupported` | a runtime resource record whose state is not the one its receipts support, read in order (the latest authoritative observation wins), or whose receipt time is outside its effect's window | fatal |\n| `runtime-resource-ttl-exceeds-bound` | a runtime resource requested for longer than its adapter's registered max-ttl | fatal |\n| `runtime-resource-unknown-adapter` | a runtime resource record whose adapter is not in the adapter registry, is not registered for the record's class, or does not implement a verb the record uses | fatal |\n| `runtime-resource-unreconciled-effect` | a runtime resource effect with outcome unknown left unreconciled, or another effect issued while one is pending | fatal |\n| `runtime-resource-unreconciled-release` | a runtime resource release counted done with no later reconcile receipt observing the resource absent | fatal |\n| `shredded-cell` | a raw `|` in the Command cell is read as a table delimiter, truncating the command and shifting every later column — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `stream-cap` | a change that adds an active stream past the per-root active-stream cap (ASSAY_STREAM_CAP) with no offsetting park — no net new streams past the cap | fatal |\n| `stream-source` | a change that adds an active stream README citing no `spec:`, or a `spec:` whose header is not `**Status:** approved` — a stream is scaffolded only from an approved spec | fatal |\n| `table-touches-no-files` | no Verify command references any path the brief's `files:` declares, so the table may prove nothing about the change (an unparseable `files:` is COULD-NOT-CHECK) | advisory |\n| `trivially-green` | a command that exits 0 whatever the tree holds — `true`, `echo ok`, a trailing `|| true`, a `git log --grep` with no count, or an existence test on a path the brief itself declares | advisory |\n| `unsubstituted-metavar` | an unsubstituted `\u003Cmetavar>` placeholder left in the Command cell, so the row cannot be run as written — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `verify-row-portability` | a Verify row hardcodes \u002Ftmp, sh\u002Fbash -c or findstr without an explicit OS marker (TMPDIR fallback is exempt) | advisory |\n| `verify-unplaced-rows` | a closure this branch makes whose `## Verify` section holds text `statusgen verifyrun` cannot place and that looks like Verify rows: a numbered row that lost its table header (a prose paragraph split the table), a table headed `Expect` with no `Command` column, or a row too short to reach `Command`. A `Failure mode | Caught by` map, a mutation map or a legend is not Verify-shaped and is not reported | fatal |\n\n## Before dispatch — mistake-proofing the brief itself\n\nThe rules above shape the brief. The three steps below **test** it, at the one pause point the flow\nalready has: *the brief is authored and the Verify table is written; nothing is dispatched yet.* All\nthree are zero-tooling — no lint reads any of them, which is exactly why they are procedure and not\nschema. The doctrine, the evidence, and the WHY live in **`docs\u002Fmistake-proofing.md`** (the public\ntool home) §§B5–B8; that document is the one home for them. What follows is only the procedure.\n\n### Pre-mortem → detection map (B5)\n\nSpend **one** prompt on prospective hindsight, about the brief you just wrote (not about the product):\n\n> *\"This work shipped and was wrong — what went wrong?\"*\n\nThen map every failure mode it names to the Verify row that would catch it:\n\n| Failure mode of the work | Caught by |\n|---|---|\n| Worker edits the source but never regenerates, so the committed output is stale | row 3 — `\u003Cregen> --check`, exit non-zero on drift |\n| The new lint rule ships but never fires; no input in the suite trips it | row 4 — deliberate bad input, expect exit 1 |\n| The version pin in the guide is written from memory and is already wrong | **no row** |\n\n**A failure mode with no row IS the finding.** Either add the row (here: dereference the pin against\nthe live release), or record in one line why it is review-only — adequacy, taste, and materiality\nstay with review by design, and saying so is an answer. What you must not do is leave the third\ncolumn blank and dispatch. The rows are the deliverable; the map is the generator that produced them,\nand may be kept in the brief beside the table or discarded once the rows land.\n\n### The dispatch checklist (B7)\n\nRun it at the pause point, **do-confirm** style — the work is already done, you are confirming the\nkiller items — in under a minute. It covers ONLY what `statusgen --lint` cannot read. It stays this\nshort: a checklist that grows becomes a document, and a document is the weakest device there is.\n\n```\nDISPATCH CHECKLIST — brief authored, before dispatch\n\n[ ] 1. Rows DISCRIMINATE. Picture a wrong-but-plausible deliverable — right shape, right\n       sections, a confident falsehood inside. Does at least one row go red on it?  (B6)\n[ ] 2. Facts dated and checkable. Every `facts:` entry is dated, or carries the command that\n       re-establishes it. Nothing written from memory.\n[ ] 3. Self-contained. Executable without reading another brief and without exploring the repo.\n[ ] 4. Risk answers match `files:`. Read the four answers against the declared paths — do those\n       paths trip a risk every answer says no to?\n[ ] 5. `gate-why` is substantive — names what about THIS brief trips the wire.\n[ ] 6. Effort and exec-tier honest. Not an L wearing an M; not `any` on work that needs `strong`.\n[ ] 7. Shared value → a FLOW row, not only a site row, and `consumers:` enumerated (rule 6).\n[ ] 8. Every new brief → `design-fit:` answered; any weight delta > 0 → `why-add` names what\n       removal was considered. New component or substantial boundary change → `layering:`\n       records justified structure; contract placement and verification follow brief-v1 §4.1\n       for any claimed boundary, new or changed (or check flat-tool behavior).\n[ ] 9. Pre-mortem run; every failure mode has a row or a recorded review-only reason.\n```\n\nItem 1 is the negative control: a Verify table that cannot fail is a green lamp wired to nothing.\nExpect the list to be wrong at first — revise it against misses you actually observe, and keep it at\nsingle digits by dropping the item that stopped earning its place, not by appending.\n\n### Comprehension probe (B8)\n\n**Optional** — reach for it when a brief's self-containedness is genuinely in doubt (dense subject,\nunfamiliar surface, a draft you compressed hard).\n\n1. Write **3–5 questions** a competent executor must be able to answer to do the work — pointed, not\n   trivia: *\"which file does the second Task step edit, and what does it currently contain?\"*,\n   *\"which Verify row fails if the generator is not rebuilt first?\"*\n2. Hand a **fresh, cheap-tier** agent the brief text ALONE — no repo, no plan, no sibling brief — and\n   have it answer. (The model-tier gate at the top of this skill bars a cheap tier from *authoring*;\n   answering a probe is not authoring. A cheap answerer is the point: it tests whether the brief\n   carries the work rather than whether the reader is strong.)\n3. A wrong or hedged answer is a defect **in the brief**. Fix the brief. Never fix it by enriching the\n   worker's dispatch prompt — the brief is what the next reader gets.\n\nA failed probe is a defect caught before it crossed the boundary, and it costs less than one failed\nworker run.\n\n## Intra-brief splits — `parallel-streams:`\n\nA large brief whose parts are genuinely independent can be worked by several concurrent workers\ninstead of one. The optional `parallel-streams:` field is how a brief declares that. **Absence is the\ndefault and always will be** — nearly every brief omits it and dispatches to exactly one worker.\n\n**Write it only when all four hold.** Otherwise leave it out; a serial brief is never wrong.\n\n1. **Effort: L, or the brief says why an M pays.** `statusgen --lint` NOTICEs a split on an S brief:\n   coordinating two workers, two claims and a recombine costs more than an S brief contains.\n2. **The parts are separable in the deliverable, not just in the file tree.** \"These touch different\n   directories\" is a statement about paths; the question is whether either half can be reviewed and\n   merged without the other.\n3. **No shard owns a shared surface.** A numbering space, a row table, a generated artifact, a module\n   graph, a pin set: two shards editing *different lines* of one of these still collide, because what\n   they contend on is not the bytes. The declared list — with the reason each surface earns its place\n   — is `sharedSurfaces` in `statusgen\u002Fshardcheck.go`; that is the one source, so do not restate the\n   list here or in a brief. Edits to those files belong to the coordinating worker, after the shards\n   land.\n4. **`statusgen shardcheck` approves it.** Declaring the field is a REQUEST. Run\n\n   ```\n   statusgen shardcheck --brief docs\u002F\u003Cstream>\u002Fbrief-\u003CNN>-\u003Cslug>.md --root .\n   ```\n\n   before dispatch; it is also runnable on a candidate split with no brief written yet\n   (`--shard engine=statusgen\u002F** --shard docs=docs\u002F\u003Cstream>\u002F**`). Three states, and only the first\n   permits a split: **0 checked-clean** · **1 checked-failed** (a named collision) ·\n   **2 could-not-check** (something it could not analyse). **1 and 2 are the same instruction —\n   dispatch ONE worker, serially.**\n\n**What the check prevents, and what it does not.** It is a precondition on the split, not a merge\ngate, and saying so plainly is the point:\n\n| Class | Prevented? | By what |\n|---|---|---|\n| Two shards editing the same file | yes | `path-overlap` — glob disjointness, checked against the real tree rather than asserted |\n| Two shards editing one numbering\u002Frow\u002Fgenerated space | yes | `shared-surface` — the surface is withdrawn from every shard |\n| Shard A changes a declaration shard B calls | yes, **within one Go package** | `symbol-coupling` — the class with no textual conflict, where both shards stay green alone and main goes red on merge |\n| Any other cross-shard file pair (non-Go, or Go across packages) | **NO** | reported as a `COVERAGE-GAP`, which forces could-not-check → serial |\n| A collision between this brief and a **different** branch | **NO** | out of scope by construction — one brief's shards are all it sees; that is merge-time detection, which is `statusgen mergecheck`'s job |\n| A branch that is stale relative to main | **NO** | merge-currency, also `mergecheck`'s |\n\nThe middle row is the one that matters. The collision this was built for was a function growing an\neighth parameter on one branch while another still called it with seven — different files, both green\nindependently. A partition that only checks paths would have looked at those two files and seen a\nclean split. So a path partition is **necessary and not sufficient**, and the gap is named rather\nthan defaulted away: any pair the checker cannot reason about produces could-not-check, and\ncould-not-check runs serial.\n\n## Phase README structure\n\n`docs\u002F\u003Cphase>\u002FREADME.md` ties the briefs together. It carries each numbered item below that\napplies under its own `##` heading (`## Briefs`, `## Critical path`, `## Dependency waves`, …).\nThe `## Dependency waves` section is never folded into the `## Critical path` section, even\nthough both show the critical path:\n\n1. **Briefs table — generated, not hand-written, once `statusgen` is wired up.**\n   The table lives between\n   `\u003C!-- statusgen:briefs:begin -->` \u002F `\u003C!-- statusgen:briefs:end -->` markers, and the README's\n   own frontmatter carries `board: generated`; `statusgen` writes the Status \u002F Verified \u002F Reviewed\n   cells from PR trailers, `verifyrun` witnesses and App approvals — a hand edit inside the markers\n   is a lint PROBLEM, the same as hand-editing `STATUS.md`. Columns:\n   `| # | Brief | Wave | Effort | Status | Verified | Reviewed |`, Status a bare lifecycle token\n   (`todo` \u002F `in-progress` \u002F `implemented` \u002F `verified` \u002F `done`, the hold token `blocked`, or the\n   off-path `withdrawn` — see \"Withdrawing a brief\" below). A\n   project not yet wired to `statusgen` hand-maintains the older\n   `| # | Brief | Wave | Status | What's deployed\u002Flanded |` shape as a stopgap — ✅ \u002F ⚠ \u002F ❌ with a\n   terse, honest \"what's actually true\" note (not aspirational) — but the moment generation is\n   live, the table is derived, never hand-edited.\n   **Withdrawing a brief.** A brief retired by a later decision, or merged into another brief,\n   is never parked as `blocked` with prose in its title. Set its Status cell to `withdrawn` (the\n   link text keeps the original title; preserve existing Verified\u002FReviewed stamps), switch its frontmatter to\n   `schema: brief-v3` (the withdrawal keys remain unknown on brief-v1\u002Fv2), and add:\n   `withdrawn-date: \"YYYY-MM-DD\"`, `superseded-by: [\"\u003Cstream>\u002F\u003CNN>\" | \"\u003Ccell>:\u003Calias>:\u003Cstream>:\u003CNN>\"]`\n   (`[]` = won't-fix) and a one-sentence `withdrawn-reason:`. Put the banner\n   `> **Withdrawn YYYY-MM-DD — superseded by \u003Csuccessor>.** \u003Creason>. Not executable.` under the H1.\n   Never delete the file or erase partial implementation, Evidence, stamps or receipts.\n   Keep the existing human decision gate: withdrawal is not permission to bypass it.\n   A withdrawn brief leaves the done ratio, Next-up and the bottleneck views; dependencies\n   remain unsatisfied until explicitly amended. Validate successor chains and review each\n   dependent's scope before re-pointing it. Use a shared, revision-bound `withdrawal-context`\n   handoff where the amendment has one; preserve observed claims with acknowledged treatment.\n   Run `statusgen withdrawal-audit` over the candidate board and an open-PR snapshot before\n   rollout, then reconcile affected implementation and evidence PRs with their owners.\n   Do not auto-close them. Upgrade readers before authoring v3; see `docs\u002Fwithdrawal-rollout.md`.\n2. **Critical path** — an ASCII chain showing the blocking order, with a one-line \"smallest unblocking\n   move\" and a warning if any tempting-but-wrong first step is a dead end.\n3. **Dependency waves** — `Wave 0: [..]  Wave 1: [..]←0  …`, plus the one-line critical path\n   (`0→8→1→2→5`).\n4. **Delivery milestones when applicable** — link the canonical records and show outcome,\n   contributor wave span, remaining work, acceptance outstanding and prerequisite holds\n   per the shared procedure. Keep generated projections in their own supported region;\n   do not add milestone rows to the Briefs table or hand-edit generated cells. Until\n   supported, put the proposed mapping in the scoping document and link it from authored prose.\n5. (Optional) \"Before starting\" prereqs and \"Shared conventions\" the briefs inherit.\n\n## When a spec is approved: file the authoring follow-on in the same motion\n\nA spec or scoping doc reaching `approved` is not yet work — nothing pulls it forward until\nbriefs exist, so an approved-but-unrouted spec is exactly the thing that gets forgotten.\nClose that edge in the landing motion: whoever lands the PR that marks a spec `approved`\nfiles, in the same motion, ONE work-ready issue on your tracker titled **\"Author briefs for\n&lt;spec path&gt; into &lt;destination stream&gt; (strong tier)\"**, naming this\n`assay:author-brief` procedure and the **strong tier** requirement in the body so it rides\nyour existing issue → work-dispatch lane to a strong-tier author. \"Same motion\" means the\nlanding PR is not complete without it: file the issue immediately after the merge, or\nreference it from the PR body before merging. Landing those briefs later flips the spec's\nstatus to `routed` in the citing PR. (An automated emitter may back this floor as a second\nlayer; the floor does not wait on one.)\n\n## Working method\n\n1. **Reconcile outcomes and inventory the work** — preserve the source's milestone promises;\n   map criterion coverage to existing briefs before adding new units. Follow the shared\n   procedure for milestone dependencies, versions and reader support. Each new unit becomes\n   a brief; keep them roughly equal-\n   sized (split anything too big — e.g. one brief doing two distinct roles becomes two).\n2. **Map dependencies** — for each unit, what must exist first (`Depends on`) and what it enables\n   (`Unblocks`). This gives the waves (Wave 0 = no deps; Wave N = depends only on \u003CN).\n3. **Derive + verify the critical path** — for each selected milestone, include its contributing\n   work, actual execution dependencies and final acceptance prerequisites. Validate mixed-graph\n   cycles and qualify paths with unknown root\u002Fevidence coverage. Verify the head per the\n   judgment-call section above; a wave summary alone is not readiness.\n4. **Write the README first** (table + path + waves), then the briefs. Cross-link both directions.\n   Minimal skeleton, one `##` heading per section:\n\n   ```markdown\n   ## Briefs\n   ## Critical path\n   ## Dependency waves\n   ```\n5. **Convert relative dates to absolute** (e.g. \"by next week\" → the actual date) — these docs\n   outlive the conversation.\n6. **Author the facts and the edges, not the cell.** Status is DERIVED, not chosen: write `why:`,\n   `depends`, `unblocks`, `gate`, `risk` accurately and let `statusgen` read the PR trailers,\n   witnesses and approvals into the Status \u002F Verified \u002F Reviewed cells.\n   A project not yet wired to `statusgen` keeps the honesty\n   rule in spirit by hand — \"⚠ Workaround\" \u002F \"Blocked by #X\" beats a green checkmark that isn't\n   true — but once generation is live, a hand-edited cell is a lint PROBLEM, not a courtesy.\n7. **Handoff to execution**: a brief is a scope-and-DoD contract, not a step-by-step plan — don't\n   write one here. When a brief (especially Effort: L) is picked up for execution, consider running\n   `superpowers:writing-plans` against that brief's Deliverables\u002FDoD to produce a bite-sized TDD plan\n   before touching code. That's a separate step, done at pickup time by whoever executes it — not\n   part of authoring the brief.\n8. **Link the authoring PR with `Authors:`, never `Brief:`.** The PR that writes brief files\n   carries exactly one link line `Authors: \u003Cstream>\u002F\u003CNN>[, \u003Cstream>\u002F\u003CNN> …]` naming every brief it\n   adds (or `Issue: #\u003CN>` when the authoring answers an issue, or `Fixes: #\u003CN>` when merging the authoring PR should close it — never for an issue carrying `needs-decision` or `human-decided`, or any decision-gate issue; those link with `Issue: #\u003CN>`). `Brief: \u003Cstream>\u002F\u003CNN>` means the\n   PR DELIVERS that brief: the dispatcher's phantom check, the planner and the derived board all\n   read it that way, so an authoring PR carrying it makes the brief it just wrote look delivered\n   the moment it merges, and the brief is never dispatched. `deskpr create` refuses a `Brief:`\n   line on a branch that only authors that brief and names the `Authors:` line to use, and refuses\n   an `Authors:` line unless the branch adds every listed brief's file and touches nothing but\n   board READMEs, brief files and changelog fragments. A PR that authors a brief AND delivers\n   work in the same change keeps `Brief:`.\n\n## Conventions to inherit\n\n- Briefs are reference docs, not narration — terse, imperative, link to the authoritative source\n  rather than duplicating it.\n- When a brief's work lands, fill its Evidence section (one row per Verify item, filled by someone who\n  did NOT implement) and update the README status row + critical path in the same change, so the plan\n  never drifts from reality.\n- **Docs regeneration**: if this project has auto-generated docs or a docs-site that must be\n  regenerated when the public-facing architecture, API surface, user flow, or operational tooling\n  changes, every such brief MUST carry a docs-regeneration item in its DoD naming the specific\n  pages\u002Fareas to update. State the project's concrete regen command in the project-level skill (or\n  the repo's instructions file).\n- **Changelog fragment**: if the target repo enforces a per-PR fragment (a `changelog\u002F` directory\n  carrying `changelog\u002FREADME.md`), LIST `changelog\u002F\u003Cslug>.md` in the brief's `files:` line — `files:`\n  is the implementer's contract, and a path absent from it will not get written. The fragment is\n  human-legible highlight bullets, not a restatement of the Task; a genuinely non-notable brief says so\n  in its `why:` and leaves the `changelog:skip` waiver to the desk or a human. Do NOT add a Verify row\n  for it — the repo's CI changelog check IS that row.\n- **Short identifiers**: keep identifiers a brief asks for (test function names especially) under\n  32 characters, and in a PR body describe a long identifier rather than quote it: the desk secret\n  scan reads any 32+ character alphanumeric run as a possible secret.\n- If executing a brief surfaces a NEW non-obvious gotcha, fold it into the repo's instructions file\n  (CLAUDE.md \u002F AGENTS.md \u002F etc.) so the next person doesn't rediscover it.\n","# Author Brief\n\n**A stream board's Briefs table is a derived, generated surface once `statusgen` is wired up** — a\nbrief states the authoring facts and the edges\n(`why:`, `depends`, `unblocks`, `gate`, `risk`); it never states the lifecycle cell, and no one\nhand-edits the generated table.\n\n## Model-tier gate (check BEFORE anything else)\n\nBrief authoring and work decomposition are **design-tier work**: this is where risk gates are\nderived, dependencies mapped, and the critical-path head verified — errors made here COMPOUND\nthrough every downstream implementer and review cycle, unlike implementation errors, which gates\ncatch. Therefore:\n\n- **If you are a fast\u002Fcheap-tier model** (a haiku-class or equivalent economy tier): **STOP.**\n  Report which model you are and tell the user to switch to a strong-tier session or escalate\n  the authoring to one. **Do not author anyway** — not for \"just a small brief set\", not as a\n  \"draft for review\", not because the user seems in a hurry. A cheap draft anchors the strong\n  model that reviews it; refusing is the cheaper path.\n- **If you are dispatching this work to a subagent**: use the most capable model available for\n  authoring\u002Fdecomposition — never an economy tier. (Implementation is the inverse: cheap\n  implementers behind strong gates.)\n\nPlanning multi-step work follows a consistent shape: a phase README that holds a **status table**, a\n**critical path**, and **dependency waves**, plus one **self-contained brief** per unit of work. The\npoint of the structure is that any single agent can pick up a brief and execute it without re-reading\nthe whole plan, and the reader can see at a glance what blocks what. **Match the phase docs already in\nthe repo — don't invent a new format.** If a project-level `author-brief` skill exists in the current\nrepo (`.claude\u002Fskills\u002Fauthor-brief\u002FSKILL.md`), read it after this skill and apply its specifics —\npaths, docs-regen commands, and local examples layer on top of this methodology.\n\n## Bind implementation pickup to current context\n\nRead the [implementation pickup procedure](..\u002F..\u002Freferences\u002Fimplementation-pickup.md).\nSpecs and briefs identify the selected target\u002Ftransition, semantic contracts and\napplicable project engineering guidelines by canonical pointer. Include their\npickup\u002Fresume assessment in the implementation handoff; record revisions, coverage\nand continue\u002Fadapt\u002Fhold rather than assume the authoring snapshot is still current.\nUse the shipped context operation where the installed version supports it (see the\nsupported version below); otherwise use the explicit manual bootstrap.\nDo not rewrite completed Evidence or let a draft proposal veto approved work.\n\n## Consume the context handoff before opening the PR\n\nWhen the spec hands over a context receipt, the brief PR answers it:\n\n1. Recheck the receipt against the current snapshot with\n   `statusgen context recheck --snapshot DIR receipt.json`. Reconcile (exit 1) or resolve the\n   hold (exit 3) before authoring on it.\n2. For each impact you dispose `amend`, write the candidate brief patch, a single-file\n   unified diff against the brief's current revision, and record the obligation in the\n   receipt's `amendments` with that patch's base revision and digest. Bind every obligation\n   in an `amendment-manifest-v1` file: owning brief, action, delivery paths, test commands,\n   active-work treatment, and the patch or follow-up brief it binds to.\n3. Run `statusgen context check --snapshot DIR receipt.json manifest.json PATCH...` and open\n   the PR only on exit 0. An `amend` obligation without a patch is a failure, not a note:\n   the receipt's own validation refuses it. Fix a refusal (exit 1) or a hold (exit 3); never\n   waive it in the PR body.\n4. The existing rules still apply: author `consumers:`, `depends:` \u002F `unblocks:` and the\n   Verify table as below. A patch never edits populated Evidence or a dated source account,\n   and a delivered brief takes a linked follow-on instead of a patch.\n5. Keep product and architecture bindings separate: carry each brief's `outcome` and\n   `outcome-revision` unchanged unless the spec moves them, and never drop the revision the\n   work was first selected against.\n\n**Supported version.** `statusgen context discover`, `statusgen context recheck` and\n`statusgen context check`, with the `context-handoff-v1` and `context-handoff-v2` receipts and\nthe `amendment-manifest-v1` file, ship in Assay v1.0.55 and later. The product-binding\nextension, `context-handoff-v3`, ships in v1.0.63 and later. `statusgen context --help`\nlists the verbs an install has.\n\n**Manual bootstrap.** When the installed statusgen lacks the verb, or lacks the\nproduct-binding extension while briefs bind `outcome-revision`, follow the\n[implementation pickup procedure](..\u002F..\u002Freferences\u002Fimplementation-pickup.md) by hand: record\nthe impacted briefs, their revisions, coverage and each amendment obligation in the handoff,\nand say that it is a manual bootstrap. Never invent a receipt or a schema version the\ninstall does not ship.\n\n## Map work to delivery milestones\n\nWhen the source defines delivery milestones or the request needs staged outcomes,\nread the [shared milestone authoring procedure](..\u002F..\u002Freferences\u002Fmilestone-authoring.md),\nincluding its support check and **Author-brief** section. Map existing\u002Fnew briefs to\n`\u003Cstream>:m-\u003CNN>` outcomes and criteria; prefer `\u003Cstream>:b-\u003CNN>` for new briefs\nwith optional `@\u003Cbrief-version>` on references; use full\n`\u003Ccell>:\u003Crepo>:\u003Cstream>:b-\u003CNN>` IDs in brief definitions. Once the pinned generator\nand required readers support brief §3.8, write colon-only forms and preserve legacy\naliases as compatibility reads. Until then use\nsupported legacy IDs for executable briefs and `b-` labels in scoping proposals;\ndisclose the limitation. Do not rewrite existing identities or Evidence. Membership is\nowned by the milestone, not a second brief field. Separate actual brief start needs\nfrom prerequisites for final outcome acceptance. If the source lacks a consequential\noutcome\u002Fscope decision, reconcile it with `author-spec` before committing affected work;\ncontinue independent authoring. Unsupported declarations remain proposals, not live YAML.\n\nFor a source using simulated dependencies, apply the shared procedure's **Outcomes\ndelivered with simulated dependencies** section. Preserve its real\u002Fsubstitute matrix\nand evidence limits in the affected briefs. Separate profile setup, contract\u002Ffake work\nand the joined real-consumer qualification where they have different owners; record\nactual producer prerequisites rather than adding another controller to hide a gap.\n\n## The most important judgment call: find the REAL head of the critical path\n\nThe critical path is the chain that, if it slips, slips everything. The recurring failure is putting\nthe wrong item at the head — sequencing work behind a step that is itself blocked. (Concrete shape of\nthe failure: a plan listed Brief B → Brief A as the unblock, but both were dead-on-arrival until an\nupstream bug — tracked elsewhere — was fixed; that bug was the true head and wasn't even in the path.)\nBefore you finalize the order: for each \"smallest unblocking move,\" **verify it actually works** —\ncheck the code\u002Fconfig\u002Fbug tracker, don't assume. If a proposed first step depends on something\nunproven, that something is the real head. Say so explicitly.\nThe requester's stated edges are a claim to check, never the evidence for the head, and the check\nis done now, while you author — never left to brief 01's Task. Do it against what you can read in\nthis session: the code, config or tracker when you have them, and always the requester's own words\nand the material you were handed. Test each stated edge there: does the step it puts first rest on\nanything that text does not establish? A run without repo access still has that text, so \"I could\nnot read the repo\" names what you did not look at; it is never the check.\nState the head in one passage, in your own prose, that names it, quotes the line the check rests\non with where it came from, and says what that line showed, in one of two forms:\n\"I verified \u003CX> is the real head by reading \u003Csource>: '\u003Cline>'; it showed \u003Cwhat>\", or\n\"\u003Csource> says '\u003Cline>', which does not establish \u003Cwhat>. \u003CX> is unproven, so it is the real\nhead: brief 01 settles it\". \"That makes 01 the head\", \"established from the stated edges\", and an\n\"unproven\" with no line read and quoted state no check.\n\n## Carry quality obligations into execution\n\nCarry applicable quality attributes and cost-model obligations from the controlling spec\ninto the brief's existing Context, Task, DoD and Verify fields. Name the affected path or\nquality scenario, operating conditions, acceptance target or unresolved decision, behaviour\nto preserve, and evidence owner and method. For performance, retain workload and budget. Cite the source; do not silently invent or relax\nbudgets. If the source omits a consequential constraint, surface the gap before committing\nwork that depends on it. Keep unrelated briefs free of boilerplate quality checklists.\n\nWhere algorithm or data structure choice is consequential, state what needs comparison and\nwhy. An unresolved choice that could invalidate downstream work needs a bounded spike ahead\nof that work, with a decision criterion and explicit dependencies. Verification should cover\napplicable resource budgets and scaling as well as functional equivalence; a faster result\nthat skips required behaviour does not pass. Reuse the spec's workload and measurement\nconditions so that the evidence answers the original requirement. Keep cross-cutting\nobligations traceable to the briefs that implement them and the brief that verifies the\nend-to-end result; passing isolated component checks is insufficient for a system target.\n\n## Brief structure (one file per unit of work)\n\n`docs\u002F\u003Cphase>\u002Fbrief-\u003CNN>-\u003Cslug>.md` (or whatever your repo's convention is — match existing phase docs).\nConceptually every brief carries the same handful of things: **Wave, Depends on, Unblocks, Effort**\n(the scheduling metadata), **Context** (why + the exact facts needed), **Read first** (required\nreading), **Deliverables** (concrete outputs), the **Interface contract** (the seam other briefs rely\non), and a **Definition of Done**.\n\nOlder brief sets narrated all of this in prose — a header line plus paragraphs. That drifts in two\nspecific ways once a plan has more than a few briefs: dependencies written as free-text arrows that\ndon't survive a rename or a re-numbering, and DoD items with no runnable check attached, so \"done\" is\nwhatever the implementer felt like asserting. **The template below is the current standard — use it\nfor every new brief.** It carries the same concepts as data (YAML frontmatter fields, tables) instead\nof only prose, so a script — or the next agent, or you in six months — can rely on the structure\nwithout re-reading the whole plan to extract it.\n\n**Start here: don't hand-type a brief — generate it.** `statusgen newbrief --stream \u003Cs> --title \u003Ct>`\nemits this skeleton with every derivable field already correct: the gate computed from the four risk\nanswers (it refuses a supplied gate, and in non-interactive mode refuses an unanswered risk question\nrather than defaulting it), the wave computed from `--depends`, the inverse `unblocks:` edge written\ninto each named dependency in the same change, and a freshness stamp from a fetch it performs. Hand-\nauthoring stays valid (fix a typo without re-running a tool); the generator just removes the fields an\nauthor can otherwise get wrong. Run `statusgen newbrief --help` for the flags.\n\n### Brief file template (data-first)\n\nSelect the identifier\u002Fschema pair before writing. With negotiated prefixed-label\nsupport, write `brief: \u003Ccell>:\u003Crepo>:\u003Cstream>:b-\u003CNN>` and keep the authored revision\nin `version:`. Dependency refs may use `:b-\u003CNN>@\u003Cversion>`; no suffix binds the selected\nsource snapshot. Use that supporting schema for references\nand dependency fields. The template below shows the currently supported legacy\n`brief-v2` pair; do not combine its schema with an unsupported `b-` identifier.\nThe unprefixed label is deprecated for new authoring after adoption, not invalid for\nlegacy consumers. Existing canonical fullIds and filenames remain unchanged (brief §3.8).\n\n```markdown\n---\nbrief: \u003Ccell>:\u003Crepo>:\u003Cstream>:\u003CNN>   # current brief-v2 fullId; board label is \u003Cstream>\u002F\u003CNN>\ntitle: \u003Cone line>\nwhy: \u003Cprose>                        # REQUIRED for every NEW brief — one to three lines a non-engineer\n                                    # could read and justify the work from. Not the WHAT (that's title\n                                    # + Task), but WHY it matters: what problem, whose pain, what the\n                                    # outcome unlocks. Quality bar: \"all 22 pages are eagerly imported\n                                    # into a single JS chunk, so every user pays a multi-MB first load;\n                                    # splitting cuts the initial payload ~5×.\" (frontend\u002F05 exemplar).\n                                    # Omit only on legacy briefs being backfilled later — statusgen\n                                    # NOTICEs a missing why: (non-fatal this phase; hard error once\n                                    # backfill lands — same pattern as gate-why).\nwave: \u003Cint>\ndepends: []                        # typed IDs only: [\"\u003Cstream>\u002F\u003CNN>\", ...] — NEVER prose arrows\nunblocks: []                       # typed IDs\neffort: S | M | L                   # closed scale — L is the CEILING (rule 12); XL is not a size,\n                                    # it is the signal to split into multiple briefs, each ≤ L.\ngate: model | human                # from the four risk questions below\nrisk: {regulatory: no, customer: no, irreversible: no, sensitive-data: no}\nissues: []                         # GH issue numbers this brief closes\nregression-of: \u003Cissue ref or sha>  # OPTIONAL — fix\u002Fbug briefs only (rule 14): the PRIOR fix this\n                                    # defect re-opens — its issue reference or its commit sha.\n                                    # Omit when no earlier fix exists (a first occurrence); never\n                                    # invent one to fill the field. No lint reads it yet: the key\n                                    # is tolerated as an unrecognised frontmatter key, not validated.\nschema: brief-v2                    # required in a v2 tree;\n                                    # --lint PROBLEMs a tree of v2 briefs missing it. Reserved,\n                                    # OPTIONAL keys parsed under brief-v2 (shape-validated only,\n                                    # gating behaviour deferred to the graph stream): id (uuid,\n                                    # minted once at authoring, never reused), supersedes ([]),\n                                    # version (int, bumped on every Task\u002FVerify edit after first\n                                    # dispatch), gates ([{on, type, reason}]), feathers ([...]).\n                                    # NO status: KEY — the lifecycle cell is DERIVED by statusgen\n                                    # from PR trailers, verifyrun witnesses and App approvals, never\n                                    # hand-asserted in frontmatter or the stream README table.\nauthored: \u003CYYYY-MM-DD> by \u003Cwho\u002Fsession>\nsources: []                        # provenance: scoping doc, finding IDs, intake IDs this derives from.\n                                    # F-\u003Cslug>\u002FI-\u003Cslug> references should be links: \"[F-ws-token-expiry](..\u002Ffindings\u002FYYYY-MM-DD-example-finding.md)\"\ngate-why: \u003Cprose>                  # REQUIRED when gate: human OR any risk answer is yes — what about THIS\n                                   # brief makes it risky and what the human is confirming (a line or two).\n                                   # Omit only when gate: model and all four risk answers are no.\ndecision-trigger: creation | start | spec   # OPTIONAL, gate: human only (absent = start) — WHEN this\n                                    # brief's self-contained decision issue is filed by the\n                                    # `\u003Cdecision-issue filer>` your repo's instructions file names\n                                    # (see rule 4; the same shape as `human:\u003Cname>` — a capability the\n                                    # adopter binds, not a tool this skill ships). A JUDGEMENT recorded\n                                    # per brief, not a global rule: `creation` = the options are already\n                                    # enumerable as authored (file on landing); `start` = the decision is\n                                    # concrete once a worker first picks the brief up — the default,\n                                    # filed by the dispatcher; `spec` = the options only become\n                                    # well-formed after the pickup-time design\u002Fplan step — the executor\n                                    # authors `## Human decision` from that step, and THAT files it.\nexec-tier: any | strong             # OPTIONAL (absent = any) — minimum execution-model tier.\n                                    # DERIVED from four complexity questions (see rule 9); any yes → strong.\nexec-tier-why: \u003Cone line>           # Recommended when exec-tier: strong — which question(s) it answered yes.\ndomain: clear | complicated | complex | chaotic   # OPTIONAL (absent = complicated) — the work's Cynefin\n                                    # domain. Governs which diagnostic to reach for: Ordered\n                                    # (clear\u002Fcomplicated) → Theory of Constraints; Complex →\n                                    # probe-sense-respond \u002F enabling constraints; Chaotic → act first,\n                                    # then sense. `statusgen --lint` PROBLEMs an unrecognized value;\n                                    # `statusgen --cynefin` shows the distribution and lists an\n                                    # absent-domain brief as Disorder. A repo that keeps a longer\n                                    # treatment binds it as `\u003Ccynefin-lens explainer>` in its own\n                                    # instructions file; the four values above are the whole contract.\nparallel-streams: []                # OPTIONAL (absent = one worker per brief — the default, and what\n                                    # nearly every brief uses). Declares that this brief's work\n                                    # decomposes into concurrent SHARDS, each scoped to file globs:\n                                    #   parallel-streams:\n                                    #     - {name: engine, files: [\"statusgen\u002F**\"]}\n                                    #     - {name: docs,   files: [\"docs\u002F\u003Cstream>\u002F**\"]}\n                                    # Presence is a REQUEST, never a permission — `statusgen shardcheck`\n                                    # decides. See \"Intra-brief splits\" below before writing it.\nconsumers: []                       # OPTIONAL — list of consumer sites when this brief changes a shared\n                                    # value. Each entry: \"\u003Cpath>: fixed-here | follow-up \u003Cstream\u002FNN> |\n                                    # out-of-scope (\u003Cwhy>)\". Absent = no shared value changed (the default).\n                                    # An AUTHORING PR declares future consumers, it does not edit them:\n                                    # route a path this brief's implementation will LATER touch as\n                                    # `follow-up \u003Cstream>\u002F\u003CNN>` at THIS brief, not `fixed-here`.\n                                    # See rule 6. Authoring convention — no lint enforces it yet\n                                    # (tracked as a planned follow-up).\n---\n\n# Brief \u003CNN> — \u003Ctitle>\n\n## Context\nfiles: \u003Cexact paths the implementer touches>\nout-of-repo files: \u003Cexact paths outside the repo (e.g. ~\u002F.claude\u002Fskills\u002F...), if any —\nsee rule 7; omit the line entirely when none>\nfacts: \u003Cthe 3-5 project facts needed — key: value, no narrative. The implementer\nmust never need to explore the repo.>\nlayering: \u003CREQUIRED for a new component\u002Fservice\u002Ftool or a substantial boundary change\n(domain extraction\u002Fdissolution, added adapter\u002Fentrypoint, logic crossing a process or\ntrust boundary); omit otherwise. One or two lines: simplest justified structure +\nmeaningful rules vs external effects (or why no extraction is warranted) + current\nreason + Task\u002FVerify references. Defaults (domain-core \u002F flat tool), contract placement,\ninterface\u002Fpurity rules, boundary checks and existing-brief amendment policy: brief-v1 §4.1.>\ndesign-fit:                  # REQUIRED on every NEW brief; its two rules: brief-v1 §4.1\n  owner: \u003Cthe one module that owns the meaning this brief touches, or n\u002Fa>\n  contract: \u003CS-\u003Cslug> row id of the semantic-owner index, or none — \u003Cwhy>>\n  retires: [\u003Cmechanism\u002Frefusal\u002Fflag\u002Ftest this brief removes>, ...]   # [] is an answer\n  weight: \u003Csigned delta per dimension: verbs, flags, refusals, rule-text lines>\n  why-add: \u003Crequired when any delta > 0: why it cannot live in the owner, and what\n           removal was considered instead; n\u002Fa otherwise>\noracle: docs\u002Finvestigations\u002F\u003Cfile>  # ONLY on a design brief a brittle `redesign` raised: its refactor\n                                    # oracle, whose five §5 rows Verify copies. Such a brief without it is not authored\n\n## Human decision\n\u003C!-- gate: human only — omit the section entirely otherwise. Lifted VERBATIM into the\n     brief's decision issue by the `\u003Cdecision-issue filer>` your repo's instructions file\n     names, so write it SELF-CONTAINED: the reader sees ONLY this text, never the brief —\n     no links, no URLs (a conforming filer refuses both), no repo paths, no references to\n     other briefs or docs. REQUIRED at authoring time when decision-trigger is creation or\n     start; when decision-trigger: spec, the EXECUTOR authors it at pickup from the\n     design\u002Fplan step and commits it into this file. -->\n\u003C3-10 lines: what is being decided and why it needs a human — enough to decide from\nthis text alone>\n\nOptions:\n1. **\u003Coption>** — \u003Cwhat it means; consequences; what happens next if chosen>\n2. **\u003Coption>** — \u003Csame>\n\nDefault if no answer: \u003Coption N after \u003Cdate\u002Fwindow>, or \"none — blocks until answered\">\n\n## Ground rules\n- NEVER git push \u002F trigger workflows \u002F run mutating kubectl. Leave commits per\n  the task instructions only.\n- Stop at `implemented` — you do not set verified\u002Fdone.\n- If anything is unclear or contradicts repo state: report NEEDS_CONTEXT, don't guess.\n\n## Task\n\u003Cwhat to build\u002Fchange — explicit steps>\n\n## Verify (executable — no prose-only DoD items)\n| # | Command | Expect |\n|---|---------|--------|\n| 1 | \u003Cliteral command> | exit 0; output contains \"\u003Cliteral>\" |\n\n## Evidence\n\u003C!-- appended at implementation time: one row per Verify item —\n     (command, exit code, output line(s) or hash, date, runner).\n     \"verified\" status in the stream README requires this section filled\n     by someone who did NOT implement. -->\n\n## Review\nGate: \u003Cmodel|human> (from frontmatter). Reviewer records verdict + date in the\nstream README table. Human gate is MANDATORY when any risk answer is yes.\n```\n\nFill every frontmatter field — an empty `sources: []` or a missing `risk:` answer is not a shortcut,\nit's a gap the reviewer will bounce back. `gate` is derived, not chosen: if any of the four risk\nquestions is `yes`, `gate` must be `human`; only when all four are `no` may `gate` be `model`.\n\n### Rules behind the template\n\n1. **Load-bearing facts live in islands, not prose.** Frontmatter, tables, and Verify rows are the\n   record; prose explains and motivates but never solely carries a fact a script or a reviewer needs.\n2. **References are typed IDs** (`stream\u002FNN`, `F-\u003Cslug>`, `I-\u003Cslug>`, legacy `F-NN`, `I-NN`, …) — no\n   fuzzy names (\"the pricing brief\"), no arrow notation (`←12a`). A typed ID survives renumbering\n   and greps cleanly. **Register IDs are letter-prefixed slugs:** `F-\u003Cslug>` for\n   findings, `I-\u003Cslug>` for intake — the slug is 10–20 chars, `[a-z0-9-]`, derived from the title\n   (example: `F-ws-token-expiry`, `I-model-mix-tiers`). Older entries use numeric IDs (`F-01`, `I-01`)\n   — those are frozen legacy and remain valid; new entries must use the slug form. **Going\n   forward, register references SHOULD be clickable links pointing at the per-entry\n   file: `[F-ws-token-expiry](..\u002Ffindings\u002FYYYY-MM-DD-example-finding.md)`. The typed ID\n   remains the primary key (link text); the link is for readers. The mapping is each entry file's\n   frontmatter `id:` — never guess slugs. Bare (unlinked) IDs stay legal but are not the preferred\n   form for new briefs.\n3. **Verify rows must be runnable by someone who didn't do the work.** A DoD\u002FVerify item with no\n   literal command and no expected exit\u002Foutput is not a DoD item — it's a hope.\n   Prefer POSIX commands (`grep -F`, forward-slash paths) for portable checks. For a genuinely\n   native Windows command, add a `Shell` column and explicitly mark the row `cmd` or `pwsh`\n   when authoring it. Never emit `findstr \u002Fc:\"…\"` with backslash paths under implicit `sh`:\n   bash reinterprets those arguments. Shell selection belongs in the reviewed Verify table;\n   an Evidence-only edit must not change the command or its shell to make a failing row pass.\n   **Say where each row runs.** When a row needs anything beyond the checkout, add an `Env`\n   column (`| # | Class | Env | Command | Expect |`): `tree`, `host[:\u003Cname>]`,\n   `service:\u003Cname>` for a named NON-PRODUCTION instance (a database, an API, a staging\n   deployment, described in `## Context`), or `cluster:\u003Cname>`. A row whose Env is not `tree`\n   is never `check:ci`, since no CI job reaches it; class it `check`. A row needing a project\n   secret or a live service is witnessed by a non-implementer where that environment exists,\n   never by the implementer's own report. The grammar and its `statusgen --lint` rules are\n   rule 45 in the brief-rules reference.\n4. **`gate: human` is mandatory when any risk answer is `yes`.** Record all four risk answers\n   (regulatory, customer, irreversible, sensitive-data), not just the resulting gate — the answers are\n   what a reviewer audits, the gate is just their conclusion. **A risk-gated brief (`gate: human` OR\n   any risk answer `yes`) MUST carry a `gate-why`** — a line or two naming what about *this* brief\n   trips the wire and what the human is actually confirming at the sign-off. A risk boolean with no\n   rationale is an unfalsifiable assertion; the verify-gate card surfaces the `gate-why` verbatim.\n   (statusgen NOTICEs a missing `gate-why` today; it becomes a hard lint error once the backfill lands.)\n   **A gate:human brief also gets a filed DECISION ISSUE.** The human gate is only real when there is\n   a concrete, self-contained tracker issue for the human to act on. Two halves, and only the first\n   ships with this methodology: `statusgen --decision-issues` emits the newly-eligible set as JSON —\n   title, `needs-decision` label, composed body, and a hidden `\u003C!-- needs-decision: \u003Cstream>\u002F\u003CNN> -->`\n   marker that dedupes, so a brief that ever had one never gets a second — and the\n   `\u003Cdecision-issue filer>` your repo's instructions file names is what actually files it. (Same shape\n   as `human:\u003Cname>`: a capability the adopting repo binds, because filing depends on the tracker.)\n   Record the resulting number back in the brief's `decision-issue:` frontmatter field; `statusgen\n   --lint` NOTICEs a gate:human brief in flight without one. WHEN it files is a per-brief judgement\n   you record as `decision-trigger:` (see the template): ask *\"could the reader pick an option from\n   what I can write down right now?\"* — yes at authoring → `creation` (author the `## Human decision`\n   section now and file as the brief lands); yes once a worker frames the work → `start` (the default;\n   the dispatcher files it, from your authored `## Human decision` section, or by deriving a fallback\n   from `gate-why`\u002F`why` when the section is absent); only after real design exists → `spec` (the\n   executor authors the section at pickup, which is what ripens the filing). The issue body must let\n   the human decide WITHOUT opening the repo — inline everything, since a conforming filer refuses\n   links and URLs in the composed body.\n5. **Provenance is required.** A brief with an empty `sources:` — no scoping doc, no finding ID, no\n   intake ID it derives from — is untraceable: no one can tell why the work exists or whether it's\n   still needed.\n6. **A brief that changes a SHARED VALUE must enumerate its consumers, and verify the FLOW — not just\n   the site.** (Cost that forced this rule: an API response field was renamed and verified green at its\n   own site — \"the endpoint returns the new field name\" — while a frontend that still read the *old*\n   field name silently broke the whole end-to-end flow; no dashboard row ever rendered. The change was\n   correct locally and broken across components.) A shared value is anything another component reads:\n   a party\u002Fidentity, an env-var NAME, a ConfigMap\u002FSecret key, a field's MEANING, a wire\u002FJSON format,\n   or a default. The trigger is cross-component: when the brief's Context `files:` references multiple\n   top-level directories (k8s\u002F, api\u002F, frontend\u002F, services\u002F, docs\u002F, docs-site\u002F, scripts\u002F, .github\u002F,\n   tools\u002F) or its Task touches a shared-surface keyword (party, identity, env-var, ConfigMap, Secret,\n   wire format, JSON format, a default).\n   - **Enumerate consumers as a structured frontmatter field**:\n     `consumers:` — an optional YAML list in the brief-v1 frontmatter (absent = the brief asserts it\n     changes no shared value). Each entry is `\u003Cpath-or-component>: \u003Crouting>` where routing is one of\n     `fixed-here` \u002F `follow-up \u003Cstream>\u002F\u003CNN>` \u002F `out-of-scope (\u003Cwhy>)`. An unlisted consumer is how\n     you strand an assumption. Project wrappers (this repo's `.claude\u002Fskills\u002Fauthor-brief\u002FSKILL.md`)\n     carry the full format spec. **The lint enforces this field** — `statusgen --consumers` is the\n     diff-aware gate (it exits non-zero on a routing claim the diff disproves), and `statusgen\n     --lint` alone already makes one class **fatal**: a `follow-up \u003Cstream>\u002F\u003CNN>` naming a brief that\n     exists in no stream README is a hard PROBLEM. The rest of the field's checks are advisory\n     NOTICEs. See the generated **Enforcement status of these rules** table below for the exact\n     per-rule status; do not restate it here (it drifts — spec §4 B9).\n   - **Never reclassify an entry to make it fit the grammar.** The routing token is a commitment\n     record — `follow-up` says work is still owed, `out-of-scope` says it isn't — so swapping one\n     for the other to satisfy a regex destroys the fact the field exists to hold. A `fixed-here`\n     wanting a `(\u003Cwhy>)` parenthetical is already legal — the grammar is tolerant of trailing detail\n     (`fixed-here (bumps the pin in the same commit)`), so write it. The remaining awkward case is a\n     `follow-up` whose target is an intake entry or an unauthored brief with no `\u003CNN>` to cite: do\n     **not** leave a bare `follow-up` on the belief nothing enforces it — a `follow-up` with no\n     target is a NOTICE, and inventing a `\u003Cstream>\u002F\u003CNN>` that does not exist is a **fatal** PROBLEM.\n     Route it truthfully as `out-of-scope (\u003Cwhy — deferred to an unauthored brief \u002F intake entry>)`,\n     which records the deferral without a false or missing target. Same reason existing briefs are\n     not backfilled: the record's truth outranks the grammar, and the grammar has a truthful form.\n   - **An authoring PR routes its OWN future deliverables to the deferred disposition, never\n     `fixed-here`.** A brief-authoring change *declares* where a change will land; it does not land\n     it — the authoring diff adds the brief and touches none of the paths the brief's own\n     implementation will later edit. `fixed-here` is a claim about THIS diff, legitimate only when\n     THIS diff actually edits the named path; asserting it on an authoring-only diff is a claim the\n     routing gate correctly disproves — \"no path it resolves to appears in the diff\" — reddening the\n     PR on its first push. So route each path this brief's own implementation will later touch to the\n     **deferred disposition**: `follow-up \u003Cstream>\u002F\u003CNN>` targeting THIS brief. The implementation\n     change that lands the path then replaces that routing with `fixed-here`, in the same change.\n     A self-targeting `follow-up` corroborates by construction — the brief's own file contains its\n     own ID, so the gate's back-reference check is satisfied the moment the brief exists. Every entry\n     still ends in exactly one of the three routing tokens; an entry with none claims nothing and can\n     corroborate nothing. Worked example — the same consumer written wrongly then rightly on an\n     authoring PR that only files the brief:\n     - wrong (asserts a fix this diff did not make; the gate disproves it): `- \"src\u002Fpipeline\u002Frouter.go: fixed-here\"`\n     - right (deferred to this brief until its implementation lands): `- \"src\u002Fpipeline\u002Frouter.go: follow-up \u003Cstream>\u002F\u003CNN> (this brief; flips to fixed-here when the implementation edits the path)\"`\n     Run your project's consumers check — the diff-aware `statusgen --consumers` gate — over the\n     branch before pushing an authoring change, so the deferred routing is confirmed locally rather\n     than by a red check in CI.\n   - **The consumers Verify row works after the merge too**: the bare\n     `statusgen --consumers --brief \u003Cstream>\u002F\u003CNN>` row re-runs on merged main with no hand-pinned\n     commits. It pins itself to the delivery and that commit's first parent, read locally: the\n     committed Evidence's `Verification-Table:` marker if there is one, otherwise the one\n     first-parent commit whose message (or a commit it merged) carries `Brief: \u003Cstream>\u002F\u003CNN>`. So\n     the implementation PR's commits carry that trailer, and a later PR for the same brief should\n     not repeat it: zero or several candidates is exit 2 (COULD-NOT-CHECK), and the remedy is an\n     explicit `--delivery \u003Cfull-sha> --base \u003Cits first parent>`.\n   - **Flow Verify row**: a shared-value brief's Verify table must carry at least one row exercising\n     the cross-component flow end-to-end (e.g. *frontend submits form → API persists → dashboard row\n     appears*), not just the changed site. Site-green does not imply flow-green. This one stays a\n     judgement call even after the planned check ships — prose Verify rows cannot be perfectly\n     classified, so the check is an advisory NOTICE, never a hard gate.\n   - **Cross-repo pairing**: when a change crosses a repo boundary in addition to a component\n     boundary, it carries BOTH a `consumers:` list (this rule) AND a sibling-repo SHA — note\n     the pair in both places.\n   This applies to **issue-driven work too** — a change to a shared value that skips a brief skips\n   this analysis entirely, which is the highest-risk path in the system. Promote such a change to a\n   brief, or at minimum run the consumer grep and record it.\n7. **Out-of-repo files must be declared — the declaration is the claim.** A brief whose scope\n   touches files outside the repo (user-level skills\u002Fsettings, e.g. `~\u002F.claude\u002F**`) MUST list the\n   exact paths in Context under `out-of-repo files:`. These files have no worktree isolation and no\n   branch-as-claim, and an edit takes effect in every running session the moment it's written —\n   review is post-hoc. So: the dispatcher **serializes** on the declaration (at most ONE such brief\n   in flight at a time, across all streams), and the implementer stages the edit as a diff in the\n   PR, applying it to the live file only as the LAST step before `implemented` — never\n   incrementally. Where the out-of-repo surface is itself a git repo (e.g. the `~\u002F.claude` stopgap\n   repo), commit the applied edit there too. A brief that touches this surface without the\n   declaration is scope drift — bounce it.\n8. **Freshness check — before authoring a fix\u002Fchange brief, verify the site isn't already\n   fixed.** Read the target site on fresh `origin\u002Fmain` (`git fetch origin && git log\n   origin\u002Fmain --oneline -- \u003Cpath>` + read the code) and record the check in the brief's\n   `sources:` as `freshness-checked \u003CYYYY-MM-DD> @ \u003Cshort-sha>`. If the deliverable is already\n   satisfied, the brief is NOT authored — resolve or re-scope the finding\u002Fissue instead. That\n   is the only reason to withhold it: a read stated to you as done counts as done (stamp its\n   sha), and a result you cannot fully see is no reason to hold — author the brief, with a first\n   Task step that re-checks. The output is the brief or a \"do not author — already fixed\"\n   determination, never a held draft or \"not authored yet\". The check is light; the waste it prevents (re-writing a fix that already landed, claiming a vuln\n   closed that was never open) is heavy.\n9. **exec-tier: complexity signals a minimum execution-model tier.** `exec-tier` is DERIVED,\n   not chosen (mirror of the gate\u002Frisk rule). The author answers four complexity questions;\n   any yes → `exec-tier: strong` (absent\u002Fany-no → `any`, the default):\n   (a) Does the Task require design decisions the facts do not fully pre-specify?\n   (b) Does correctness depend on cross-component\u002Fcross-artifact reasoning (shared values,\n       end-to-end flows, sweeping a pattern across sites)?\n   (c) Is it code where a subtle implementation error survives the brief's own tests (auth,\n       funds, concurrency, safety plumbing)?\n   (d) Is this a design brief raised by an error-class trigger? (A recurring defect class, not one symptom, has accrued enough counted instances or merged fixes to be owed a design.)\n   `strong` SHOULD carry a one-line `exec-tier-why` naming the question(s) (lint: PROBLEM if unknown, NOTICE if missing).\n   **Honest limitation:** statusgen never verifies which model actually ran — pickup-side\n   compliance is honor-system self-report; the enforced surface is desk dispatch + the review gate.\n10. **Defense in depth is the DEFAULT design posture.** When a brief's Task has design\n    alternatives, prefer the LAYERED one: no single control standing between a fault and the\n    damage. **Mandatory for core-system briefs** — each project layer defines what \"core system\"\n    means for its own repo (typically: anything touching money, auth\u002Fidentity, or an irreversible\n    external write); recommended everywhere else. Three authoring obligations:\n    - (a) **Single-point-of-failure note in `## Context`** — one or two lines answering \"what is the\n      ONE control this design depends on?\". If the honest answer is a single thing, the design needs\n      another layer BEFORE the brief is authored, not after a reviewer bounces it.\n    - (b) **At least two INDEPENDENT layers in the Task**, where feasible. **The independence test:\n      two layers are independent only if they fail for DIFFERENT reasons in DIFFERENT components.**\n      The same check pasted into N call sites is one layer — it fails once and everywhere. A typical\n      independent set: an enforcement point at the trust boundary (the system itself rejects the\n      operation) + validation in the service that calls it + an out-of-band monitor\u002Freconciler that\n      notices when both were bypassed.\n    - (c) **Verify rows exercising each layer INDEPENDENTLY, including at least one NEGATIVE-PATH\n      row** — break or bypass the upper layer and prove the lower one still catches the fault. A\n      Verify table that only walks the happy path through every layer at once has verified exactly\n      one layer: the outermost.\n    **Boundary — this is NOT \"add more assertions\".** Layers are DESIGNED controls at DISTINCT trust\n    boundaries, not one check duplicated for reassurance. Where a project's constraints say invariant\n    checks are closed-form math rather than runtime assertions, that constraint stands; the\n    independence test in (b) is what distinguishes a real layer from assert-spam. Adding a \"layer\"\n    that fails for the same reason as the layer it backs up costs review time and buys nothing.\n11. **A brief whose deliverable makes checkable factual claims needs at least one DEREFERENCING\n    Verify row, not only presence\u002Fformatting checks.** Two rows can both go green without proving\n    the same thing:\n    - **Presence\u002Fformatting** — `grep -c`, `wc -w`, \"section exists\" — proves the deliverable is\n      well-formed. It CANNOT fail on a wrong-but-well-formed document: a confident falsehood sitting\n      in the right section, at the right length, passes exactly like the truth.\n    - **Dereferencing** — fetch a link and check what it actually serves; run a documented command\n      and compare its real output\u002Fexit code to the specific claim; check a documented ID or property\n      against the live system it names. This CAN fail on a wrong-but-well-formed document.\n\n    Rule 3 above (\"Verify rows must be runnable by someone who didn't do the work\") still stands —\n    this rule doesn't relax it. It says that when the deliverable itself carries checkable facts (a\n    setup\u002Fconfig guide, a market or competitive analysis, a spec — anything asserting \"X is true\"\n    that a reader could act on), a presence gate can't be the *whole* mechanical floor: at least one\n    row has to be capable of catching a wrong claim, not just a missing section. Skip this only for\n    genuinely presence-only deliverables with no factual claim to dereference (a docs reformat, a\n    template scaffold) — manufacturing a dereferencing row against a deliverable with nothing to\n    check is its own kind of checkmark-DoD.\n\n    It is a DIFFERENT AXIS from the row-runner rules and the unfailable-row lint: those read the\n    command TEXT and are enforced by `statusgen --lint`; this one needs to know what the deliverable\n    claims, so no lint enforces it. A row can pass every one of them and still measure only presence.\n\n    The triggering evidence (an all-presence table that passed 8\u002F8 on a factually wrong guide) and\n    why a link-resolution lint would cover only link-shaped dereferencing are in the brief-rules\n    reference, rule 43. Don't wait for such a lint to exist before writing the row this rule asks for.\n12. **Sizing limits — L is the hard ceiling; prefer M for strong-tier \u002F risk-gated work.** The\n    `effort:` scale is closed at S | M | L, and the top of it is a rule, not a convention: **no\n    brief may be larger than L. A unit of work that would be XL (or bigger) MUST be split into\n    multiple briefs, each ≤ L**, sequenced with `depends:`\u002F`unblocks:` (or sharded via\n    `parallel-streams:` when the intra-brief split rules below hold). XL is only ever a\n    signal-to-split, never an emitted size — `statusgen --lint` already PROBLEMs any effort\n    outside S\u002FM\u002FL.\n\n    **Preference, explicitly NOT a rule:** when a brief is `exec-tier: strong` OR risk-gated\n    (`gate: human` \u002F any risk answer `yes`), **prefer M as the ceiling** — lean toward splitting\n    an L into Ms. An L remains permitted for strong\u002Fhigh-risk work when there is a real reason to\n    keep the unit whole; the default lean is M-or-smaller. Rationale: strong-tier work\n    (cross-artifact correctness) and high-risk work are exactly where a large brief most degrades\n    review quality, so smaller, more-reviewable units are preferred there — the review gate is\n    the control the risk answers lean on, and it only bites on a diff a reviewer can actually\n    hold.\n\n13. **A Verify row MAY declare an OBLIGATION alongside its class — the prose MUSTs of rules 6, 10 and\n    11 carried as typed data.** The optional `Class` cell (rule 45 in the brief-rules reference) answers\n    WHO EXECUTES a row (`check:ci` \u002F `check` \u002F `gate:model` \u002F `gate:human`). A second, orthogonal axis\n    answers WHAT OBLIGATION the row discharges, written as `+`-prefixed tokens appended to the same\n    cell — a **compound cell**, `\u003Cexecution> +\u003Cobligation> …`, so the table's column set is unchanged\n    and a legacy column-less table is untouched. Four obligation values, each an example row:\n\n    | Obligation | What the row must do | Example `Class` cell |\n    |---|---|---|\n    | `+mutation` | break the thing the change guards and prove the guard reddens (rule 10c, spec D1) | `check:ci +mutation` |\n    | `+flow` | exercise the cross-component path end to end, not just the changed site (rule 6) | `check +flow` |\n    | `+dereference` | resolve a claim rather than count its presence (rule 11) | `gate:model +dereference` |\n    | `+neighbour` | exercise a sibling site the change did not touch | `check +neighbour` |\n\n    `statusgen --lint` derives which obligation a brief OWES from the shape of its change (the branch\n    diff, its declared paths, its task prose) and reports one owed-but-absent — it checks the PRESENCE\n    of the typed row, never its ADEQUACY, which stays the reviewer's call (spec §3 D7). An unknown\n    `+token` is a hard PROBLEM, exactly as an unknown class is. **Enforcement status:** the derivation\n    lands advisory (a NOTICE); `+mutation`'s promotion to a hard gate is a follow-up. This status line\n    is deliberately minimal — the authoritative per-obligation enforcement state is generated, not\n    hand-copied here (a hand-written status becomes the next stale second copy).\n\n14. **A fix\u002Fbug brief closes the defect CLASS, not the one instance.** A defect repaired at one\n    site comes back at another when the fix closed the instance and left the class open: a second\n    caller reaches the same hazardous primitive by a different path, a test stub hides it, and the\n    regression reads as a new bug. So a brief whose Task fixes a defect carries two things beyond\n    the ordinary template:\n    - **`regression-of:`** in the frontmatter, when an earlier fix for the same defect exists —\n      that fix's issue reference or commit sha. It records that the class has escaped once\n      already, and it points the reviewer at the guard that failed to hold. Omit it for a first\n      occurrence; never invent a value to fill it. Run the freshness check (rule 8) over the\n      earlier fix's site as well as the new one.\n    - **A MANDATORY class-guard Verify row.** Name the class in `## Context` `facts:` — the shape\n      every instance shares, not the one line that failed — and add a row whose check covers EVERY\n      site the defect can recur at, not only the reported one. The model is an allow-list\n      structural test: enumerate every caller of the hazardous primitive (`exampleRawToken()`) and\n      fail on any caller outside a short committed allow-list (`exampleSafeToken()`). A lint rule, a\n      type that makes the hazardous call unrepresentable, or a single choke point serves equally. A\n      test of the reported instance may sit beside it; it never replaces it.\n    - **The class-guard row is FAIL-FIRST.** Before the fix lands, the implementer plants a\n      deliberate SECOND instance of the defect at a site the fix does not touch and shows the row\n      RED on it; the Evidence records that red run, naming the planted site, beside the green one.\n      A guard shown red only against the reported instance proves it sees that instance, which the\n      instance test already did. Where the guard's matcher could silently stop matching, keep the\n      plant as a committed positive-control fixture the guard must flag, so the red stays\n      re-runnable after merge. Tag the row `+mutation` in its `Class` cell (rule 13) — breaking\n      the guarded property and proving the guard reddens is exactly that obligation.\n\n    Worked example (neutral names) — the frontmatter line and the two class-guard rows:\n\n    ```markdown\n    regression-of: \u003Cthe earlier fix's issue reference or commit sha>\n\n    | # | Command | Expect | Class |\n    |---|---------|--------|-------|\n    | 2 | `go test .\u002Fexamplepkg\u002F -run 'TestExampleRawTokenAllowList$' -count=1` | exit 0 — every caller of `exampleRawToken()` in the tree is on the allow-list | check:ci +mutation |\n    | 3 | `go test .\u002Fexamplepkg\u002F -run 'TestExampleRawTokenAllowListFlagsPlant$' -count=1` | exit 0 — the guard, run over `testdata\u002Fexample-plant\u002F` (one planted caller outside the allow-list), reports that caller | check:ci +mutation |\n    ```\n\n    Row 3 is the positive control that keeps the fail-first re-runnable; the pre-fix red run with a\n    planted second caller in the real tree goes in the Evidence. **Enforcement status:** no lint\n    enforces this rule yet — `regression-of:` is tolerated as an unrecognised frontmatter key rather\n    than validated, and whether a class-guard row is present and truly covers the class is the\n    reviewer's call. A defect with no mechanically checkable shape (a one-off logic error nothing\n    else can repeat) states that in `## Context`, with the reason, in place of the row; the reviewer\n    weighs it, and it is not available for a defect that has already reached a second site.\n\nKeep a brief self-contained: if executing it requires knowledge from another brief, either link it\nunder \"Read first\" \u002F `facts:` or state the dependency in `depends:` — never assume the reader has the\nwhole plan in context.\n\n## Enforcement status of these rules (generated — do not hand-edit)\n\nRegenerate with `statusgen enforcement-status` piped through the skillslint sync\n(`go run .\u002Ftools\u002Fskillslint --sync`); the byte-diff gate in the skillslint CI job fails\nif this block drifts from the lint's rule registry. Hand-edit it and the gate reddens.\n\nThis table reports what `statusgen --lint` enforces, **not** what the methodology\nrequires. A rule enforced only by a tool outside the lint — a CI workflow, a desk-side\nguard — reads as `not enforced` here unless it is registered. The three statuses are\nexact: **fatal** is a `--lint` PROBLEM that makes the run exit non-zero; **advisory** is\na NOTICE that is printed but never gates; **not enforced** means no lint rule checks it —\nit is an authoring convention only.\n\n| rule | what it checks | status |\n| --- | --- | --- |\n| `bre-alternation` | a pipe in a basic-regex grep pattern (no `-E`\u002F`-P`) is an ordinary character, so the pattern matches the Verify row itself — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `cmd-marker-ambiguous` | a Command cell carrying more than one `cmd:`-marked code span, so which command the row names is ambiguous | advisory |\n| `cmd-marker-not-honoured` | a `cmd:` span verifyrun ignores because the rendered cell may not show it as code (backslash-escaped backticks; raw HTML, a link, an image, a dollar in any spelling or a character reference in the cell's prose; a span fused to the text before it; or a marker not set apart by whitespace), so the row runs its first span | advisory |\n| `cmd-marker-overrides-command` | a `cmd:` marker that replaces a first code span which reads as a command itself (multi-word), so the row runs something other than the span a reader sees first | advisory |\n| `cmd-marker-vacuous` | a `cmd:`-marked command that cannot fail (`true`, `:`, `exit 0`, a lone `echo`), so the row passes whatever the tree holds | advisory |\n| `consumers-flow-verify-row` | that a shared-value brief's Verify table carries at least one row exercising the cross-component flow end-to-end — a judgement call no lint decides | not enforced |\n| `consumers-followup-missing-brief` | a `consumers: follow-up \u003Cstream>\u002F\u003CNN>` whose target is not a brief in any stream README — the routing claim is false | fatal |\n| `consumers-followup-no-target` | a `follow-up` routing that names no `\u003Cstream>\u002F\u003CNN>` target — a deferral with no holder | advisory |\n| `consumers-followup-one-way` | a `follow-up` target that exists but never references the deferring brief back — the coverage claim is one-way | advisory |\n| `consumers-missing-list` | a brief whose prose reads as changing a shared surface but enumerates no `consumers:` (a heuristic prompt, never a verdict) | advisory |\n| `consumers-no-verify-row` | a brief carrying `consumers:` but no Verify row that runs `statusgen --consumers` to corroborate the routing | advisory |\n| `consumers-out-of-scope-no-reason` | an `out-of-scope` routing with no substantive reason for the reviewer who must weigh the exclusion | advisory |\n| `consumers-prose` | a `consumers:` written as a prose paragraph rather than a routed list, so nothing can corroborate it | advisory |\n| `consumers-unrouted` | a `consumers:` entry that names no routing token (`fixed-here` \u002F `follow-up` \u002F `out-of-scope`) | advisory |\n| `eligibility-could-not-check` | a gates: edge (or unsatisfied depends:) the eligibility evaluator could not resolve offline — an unregistered\u002Funpublished alias, an absent sibling checkout, or a forge-backed target — which HOLDS the brief out of Next-up | advisory |\n| `ere-literal-pipe` | a `\\|` inside a `grep -E` pattern is a literal pipe, not alternation, so the row matches almost nothing and passes blind — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `fail-first-closure` | a risk-bearing Verify row of a closure this branch makes (verified\u002Fdone, and not so at the merge-base) has no current fail-first witness, is unproven or non-discriminating, was run for an earlier command, or names a base that is not the merge-base or an ancestor of it that already holds the brief at its current path (only the merge-base itself when the brief was created or renamed on the branch) (docs\u002Fverify-row-strength.md) | fatal |\n| `gnu-only` | a GNU-only shell construct that fails on the BSD\u002FmacOS userland a reviewer may run the row on — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `gorun-exit` | a `go run` in the Command cell flattens the program's exit code, so a non-zero result reads as success — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `gotest-run-vacuous` | a `go test -run` selector with no `--- PASS` assertion in the same command, so the row passes whether or not the named test exists, is built, or was ever renamed away (open briefs only; a closed brief's rows are summarised, not individually flagged) | advisory |\n| `grep-pipe-zero-count` | a grep feeding a later pipeline stage under a zero-count Expect fails on its own success path under pipefail (grep exits 1 on no match), and passes on exit status alone when a match is present | advisory |\n| `grep-zero-count` | a `grep -c` whose pass bar is satisfied by a zero count measures nothing — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `moving-ref` | a diff base pinned to a moving ref (a branch name, not a SHA) makes the row's result drift under it — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `no-output-assertion` | an Expect of exactly `exit 0` on a command whose output nobody asserts on (use `exit 0; output contains \"\u003Cliteral>\"` or `exit 0; \u003CN> lines`) | advisory |\n| `pattern-containment-set` | a containment list that is not exactly the five containment values, on a node that is not an effect-free human decision, with a pre-selected default, or a signal pattern without exactly one containment node — containment could become an automated act | fatal |\n| `pattern-culprit-record` | a culprit record whose fields are not the fixed six, or whose default or withheld class is not judgement, or whose withheld file is not null — a withheld or unresolved culprit could route to the mechanical exit | fatal |\n| `pattern-decision-options` | `options`, `default` or `containment` on a node that is not a decision, a default outside the node's options, or a duplicate option — a non-decision node would carry a person's choice | fatal |\n| `pattern-effect-exceeds-role` | a pattern node whose declared effect kind is not permitted for its role, per the role-to-effect-kind table in spec\u002Fworkflow-pattern-v1.md §7 — a generated instance would carry a permission its role does not hold | fatal |\n| `pattern-effect-target-not-owned` | a non-effect-kind pattern node declares an effect whose target is not among its own outputs — a node claiming a consequence outside its own declared output boundary | fatal |\n| `pattern-join-not-check` | a pattern's `join` names a node whose kind is not `check` — the pattern has no independent integration check | fatal |\n| `pattern-judgement-no-default` | a signal class whose judgement exit is not a human decision with options and a default among them — the person closing the decision gets no recommended next step | fatal |\n| `pattern-remedy-set-open` | a signal class listing a remedy outside the closed remedy set, a mechanical exit with no remedy or not kind effect, or no `.github\u002F` refused-path while any node declares effects — an automated fix could touch gate configuration | fatal |\n| `pattern-replay-not-mandatory` | a signal pattern with no node producing the reserved `replay-fixture` artifact, or a producer that owes it under no mandatory command claim — a fix could land with no reproduction of the failure | fatal |\n| `pattern-review-same-role` | a pattern node whose evidence includes a review claim but shares its role with the node that produced its input — the implementer\u003C->reviewer separation is not machine-checked | fatal |\n| `pattern-risk-input-missing-verdict` | a pattern's `risk-input` omits one of the four risk-class verdicts (low\u002Fstandard\u002Felevated\u002Fhuman) — an instance of that risk class has no declared mandatory gates | fatal |\n| `pattern-signal-admission-open` | a signal class admitting a trigger outside schedule\u002Fpush, push from a ref other than refs\u002Fheads\u002Fmain, or no admission at all while any node in the pattern declares effects — a run of unchecked provenance could reach an automated effect | fatal |\n| `pattern-signal-input` | a signal pattern whose signal input is malformed or off an entry node, whose input classes differ from the declared signal classes, that has a node unreachable from the signal input, or whose class names a missing exit or an incomplete gate-red record — a class could escape the rules that range over the input | fatal |\n| `pipeline-exit-sunk` | a shell pipeline whose real exit status is sunk by a later stage, so the row cannot fail — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `prose-led-command` | a Command cell whose lift — the `cmd:`-marked command, else the first code span — is not a command: a prose mention (a file, a path, an `owner\u002Frepo`, a code identifier, a word ahead of the real command), a shell comment, or a shell no-op (`:`, a cmd.exe `REM` \u002F `::` remark); verifyrun and the desk executors record the row could-not-run. Fix a mention by marking the real command with a `cmd:` code span; fix a comment or no-op, marked or not, by replacing it with the command that performs the check | advisory |\n| `rE2-literal-pipe` | a `\\|` inside a `go test -run`\u002F`-bench` selector is a literal pipe in RE2, not alternation — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `runtime-resource-claim-without-principal` | a runtime resource claim (or a release of a claimed resource) not authorized by a human or organization principal, or ownership no succeeded claim backs | fatal |\n| `runtime-resource-credential-not-rotated` | a runtime resource claim with no recorded rotation of the agent credential, or a release with no recorded revocation | fatal |\n| `runtime-resource-credential-outlives-ttl` | a runtime resource whose agent credential lifetime exceeds the resource's ttl | fatal |\n| `runtime-resource-expired` | an unclaimed runtime resource past its ttl, or a claimed one past its retention, still recorded as live | fatal |\n| `runtime-resource-receipt-missing` | a runtime resource effect recorded succeeded with no adapter receipt | fatal |\n| `runtime-resource-schema-violation` | a runtime resource record that breaks runtime-resource-v1's shape, repeats an effect id, or carries a ttl, lifetime or timestamp the rules cannot evaluate | fatal |\n| `runtime-resource-state-unsupported` | a runtime resource record whose state is not the one its receipts support, read in order (the latest authoritative observation wins), or whose receipt time is outside its effect's window | fatal |\n| `runtime-resource-ttl-exceeds-bound` | a runtime resource requested for longer than its adapter's registered max-ttl | fatal |\n| `runtime-resource-unknown-adapter` | a runtime resource record whose adapter is not in the adapter registry, is not registered for the record's class, or does not implement a verb the record uses | fatal |\n| `runtime-resource-unreconciled-effect` | a runtime resource effect with outcome unknown left unreconciled, or another effect issued while one is pending | fatal |\n| `runtime-resource-unreconciled-release` | a runtime resource release counted done with no later reconcile receipt observing the resource absent | fatal |\n| `shredded-cell` | a raw `|` in the Command cell is read as a table delimiter, truncating the command and shifting every later column — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `stream-cap` | a change that adds an active stream past the per-root active-stream cap (ASSAY_STREAM_CAP) with no offsetting park — no net new streams past the cap | fatal |\n| `stream-source` | a change that adds an active stream README citing no `spec:`, or a `spec:` whose header is not `**Status:** approved` — a stream is scaffolded only from an approved spec | fatal |\n| `table-touches-no-files` | no Verify command references any path the brief's `files:` declares, so the table may prove nothing about the change (an unparseable `files:` is COULD-NOT-CHECK) | advisory |\n| `trivially-green` | a command that exits 0 whatever the tree holds — `true`, `echo ok`, a trailing `|| true`, a `git log --grep` with no count, or an existence test on a path the brief itself declares | advisory |\n| `unsubstituted-metavar` | an unsubstituted `\u003Cmetavar>` placeholder left in the Command cell, so the row cannot be run as written — fatal for a closure this branch makes (verified\u002Fdone, and not so at the merge-base); a NOTICE on every other brief | fatal |\n| `verify-row-portability` | a Verify row hardcodes \u002Ftmp, sh\u002Fbash -c or findstr without an explicit OS marker (TMPDIR fallback is exempt) | advisory |\n| `verify-unplaced-rows` | a closure this branch makes whose `## Verify` section holds text `statusgen verifyrun` cannot place and that looks like Verify rows: a numbered row that lost its table header (a prose paragraph split the table), a table headed `Expect` with no `Command` column, or a row too short to reach `Command`. A `Failure mode | Caught by` map, a mutation map or a legend is not Verify-shaped and is not reported | fatal |\n\n## Before dispatch — mistake-proofing the brief itself\n\nThe rules above shape the brief. The three steps below **test** it, at the one pause point the flow\nalready has: *the brief is authored and the Verify table is written; nothing is dispatched yet.* All\nthree are zero-tooling — no lint reads any of them, which is exactly why they are procedure and not\nschema. The doctrine, the evidence, and the WHY live in **`docs\u002Fmistake-proofing.md`** (the public\ntool home) §§B5–B8; that document is the one home for them. What follows is only the procedure.\n\n### Pre-mortem → detection map (B5)\n\nSpend **one** prompt on prospective hindsight, about the brief you just wrote (not about the product):\n\n> *\"This work shipped and was wrong — what went wrong?\"*\n\nThen map every failure mode it names to the Verify row that would catch it:\n\n| Failure mode of the work | Caught by |\n|---|---|\n| Worker edits the source but never regenerates, so the committed output is stale | row 3 — `\u003Cregen> --check`, exit non-zero on drift |\n| The new lint rule ships but never fires; no input in the suite trips it | row 4 — deliberate bad input, expect exit 1 |\n| The version pin in the guide is written from memory and is already wrong | **no row** |\n\n**A failure mode with no row IS the finding.** Either add the row (here: dereference the pin against\nthe live release), or record in one line why it is review-only — adequacy, taste, and materiality\nstay with review by design, and saying so is an answer. What you must not do is leave the third\ncolumn blank and dispatch. The rows are the deliverable; the map is the generator that produced them,\nand may be kept in the brief beside the table or discarded once the rows land.\n\n### The dispatch checklist (B7)\n\nRun it at the pause point, **do-confirm** style — the work is already done, you are confirming the\nkiller items — in under a minute. It covers ONLY what `statusgen --lint` cannot read. It stays this\nshort: a checklist that grows becomes a document, and a document is the weakest device there is.\n\n```\nDISPATCH CHECKLIST — brief authored, before dispatch\n\n[ ] 1. Rows DISCRIMINATE. Picture a wrong-but-plausible deliverable — right shape, right\n       sections, a confident falsehood inside. Does at least one row go red on it?  (B6)\n[ ] 2. Facts dated and checkable. Every `facts:` entry is dated, or carries the command that\n       re-establishes it. Nothing written from memory.\n[ ] 3. Self-contained. Executable without reading another brief and without exploring the repo.\n[ ] 4. Risk answers match `files:`. Read the four answers against the declared paths — do those\n       paths trip a risk every answer says no to?\n[ ] 5. `gate-why` is substantive — names what about THIS brief trips the wire.\n[ ] 6. Effort and exec-tier honest. Not an L wearing an M; not `any` on work that needs `strong`.\n[ ] 7. Shared value → a FLOW row, not only a site row, and `consumers:` enumerated (rule 6).\n[ ] 8. Every new brief → `design-fit:` answered; any weight delta > 0 → `why-add` names what\n       removal was considered. New component or substantial boundary change → `layering:`\n       records justified structure; contract placement and verification follow brief-v1 §4.1\n       for any claimed boundary, new or changed (or check flat-tool behavior).\n[ ] 9. Pre-mortem run; every failure mode has a row or a recorded review-only reason.\n```\n\nItem 1 is the negative control: a Verify table that cannot fail is a green lamp wired to nothing.\nExpect the list to be wrong at first — revise it against misses you actually observe, and keep it at\nsingle digits by dropping the item that stopped earning its place, not by appending.\n\n### Comprehension probe (B8)\n\n**Optional** — reach for it when a brief's self-containedness is genuinely in doubt (dense subject,\nunfamiliar surface, a draft you compressed hard).\n\n1. Write **3–5 questions** a competent executor must be able to answer to do the work — pointed, not\n   trivia: *\"which file does the second Task step edit, and what does it currently contain?\"*,\n   *\"which Verify row fails if the generator is not rebuilt first?\"*\n2. Hand a **fresh, cheap-tier** agent the brief text ALONE — no repo, no plan, no sibling brief — and\n   have it answer. (The model-tier gate at the top of this skill bars a cheap tier from *authoring*;\n   answering a probe is not authoring. A cheap answerer is the point: it tests whether the brief\n   carries the work rather than whether the reader is strong.)\n3. A wrong or hedged answer is a defect **in the brief**. Fix the brief. Never fix it by enriching the\n   worker's dispatch prompt — the brief is what the next reader gets.\n\nA failed probe is a defect caught before it crossed the boundary, and it costs less than one failed\nworker run.\n\n## Intra-brief splits — `parallel-streams:`\n\nA large brief whose parts are genuinely independent can be worked by several concurrent workers\ninstead of one. The optional `parallel-streams:` field is how a brief declares that. **Absence is the\ndefault and always will be** — nearly every brief omits it and dispatches to exactly one worker.\n\n**Write it only when all four hold.** Otherwise leave it out; a serial brief is never wrong.\n\n1. **Effort: L, or the brief says why an M pays.** `statusgen --lint` NOTICEs a split on an S brief:\n   coordinating two workers, two claims and a recombine costs more than an S brief contains.\n2. **The parts are separable in the deliverable, not just in the file tree.** \"These touch different\n   directories\" is a statement about paths; the question is whether either half can be reviewed and\n   merged without the other.\n3. **No shard owns a shared surface.** A numbering space, a row table, a generated artifact, a module\n   graph, a pin set: two shards editing *different lines* of one of these still collide, because what\n   they contend on is not the bytes. The declared list — with the reason each surface earns its place\n   — is `sharedSurfaces` in `statusgen\u002Fshardcheck.go`; that is the one source, so do not restate the\n   list here or in a brief. Edits to those files belong to the coordinating worker, after the shards\n   land.\n4. **`statusgen shardcheck` approves it.** Declaring the field is a REQUEST. Run\n\n   ```\n   statusgen shardcheck --brief docs\u002F\u003Cstream>\u002Fbrief-\u003CNN>-\u003Cslug>.md --root .\n   ```\n\n   before dispatch; it is also runnable on a candidate split with no brief written yet\n   (`--shard engine=statusgen\u002F** --shard docs=docs\u002F\u003Cstream>\u002F**`). Three states, and only the first\n   permits a split: **0 checked-clean** · **1 checked-failed** (a named collision) ·\n   **2 could-not-check** (something it could not analyse). **1 and 2 are the same instruction —\n   dispatch ONE worker, serially.**\n\n**What the check prevents, and what it does not.** It is a precondition on the split, not a merge\ngate, and saying so plainly is the point:\n\n| Class | Prevented? | By what |\n|---|---|---|\n| Two shards editing the same file | yes | `path-overlap` — glob disjointness, checked against the real tree rather than asserted |\n| Two shards editing one numbering\u002Frow\u002Fgenerated space | yes | `shared-surface` — the surface is withdrawn from every shard |\n| Shard A changes a declaration shard B calls | yes, **within one Go package** | `symbol-coupling` — the class with no textual conflict, where both shards stay green alone and main goes red on merge |\n| Any other cross-shard file pair (non-Go, or Go across packages) | **NO** | reported as a `COVERAGE-GAP`, which forces could-not-check → serial |\n| A collision between this brief and a **different** branch | **NO** | out of scope by construction — one brief's shards are all it sees; that is merge-time detection, which is `statusgen mergecheck`'s job |\n| A branch that is stale relative to main | **NO** | merge-currency, also `mergecheck`'s |\n\nThe middle row is the one that matters. The collision this was built for was a function growing an\neighth parameter on one branch while another still called it with seven — different files, both green\nindependently. A partition that only checks paths would have looked at those two files and seen a\nclean split. So a path partition is **necessary and not sufficient**, and the gap is named rather\nthan defaulted away: any pair the checker cannot reason about produces could-not-check, and\ncould-not-check runs serial.\n\n## Phase README structure\n\n`docs\u002F\u003Cphase>\u002FREADME.md` ties the briefs together. It carries each numbered item below that\napplies under its own `##` heading (`## Briefs`, `## Critical path`, `## Dependency waves`, …).\nThe `## Dependency waves` section is never folded into the `## Critical path` section, even\nthough both show the critical path:\n\n1. **Briefs table — generated, not hand-written, once `statusgen` is wired up.**\n   The table lives between\n   `\u003C!-- statusgen:briefs:begin -->` \u002F `\u003C!-- statusgen:briefs:end -->` markers, and the README's\n   own frontmatter carries `board: generated`; `statusgen` writes the Status \u002F Verified \u002F Reviewed\n   cells from PR trailers, `verifyrun` witnesses and App approvals — a hand edit inside the markers\n   is a lint PROBLEM, the same as hand-editing `STATUS.md`. Columns:\n   `| # | Brief | Wave | Effort | Status | Verified | Reviewed |`, Status a bare lifecycle token\n   (`todo` \u002F `in-progress` \u002F `implemented` \u002F `verified` \u002F `done`, the hold token `blocked`, or the\n   off-path `withdrawn` — see \"Withdrawing a brief\" below). A\n   project not yet wired to `statusgen` hand-maintains the older\n   `| # | Brief | Wave | Status | What's deployed\u002Flanded |` shape as a stopgap — ✅ \u002F ⚠ \u002F ❌ with a\n   terse, honest \"what's actually true\" note (not aspirational) — but the moment generation is\n   live, the table is derived, never hand-edited.\n   **Withdrawing a brief.** A brief retired by a later decision, or merged into another brief,\n   is never parked as `blocked` with prose in its title. Set its Status cell to `withdrawn` (the\n   link text keeps the original title; preserve existing Verified\u002FReviewed stamps), switch its frontmatter to\n   `schema: brief-v3` (the withdrawal keys remain unknown on brief-v1\u002Fv2), and add:\n   `withdrawn-date: \"YYYY-MM-DD\"`, `superseded-by: [\"\u003Cstream>\u002F\u003CNN>\" | \"\u003Ccell>:\u003Calias>:\u003Cstream>:\u003CNN>\"]`\n   (`[]` = won't-fix) and a one-sentence `withdrawn-reason:`. Put the banner\n   `> **Withdrawn YYYY-MM-DD — superseded by \u003Csuccessor>.** \u003Creason>. Not executable.` under the H1.\n   Never delete the file or erase partial implementation, Evidence, stamps or receipts.\n   Keep the existing human decision gate: withdrawal is not permission to bypass it.\n   A withdrawn brief leaves the done ratio, Next-up and the bottleneck views; dependencies\n   remain unsatisfied until explicitly amended. Validate successor chains and review each\n   dependent's scope before re-pointing it. Use a shared, revision-bound `withdrawal-context`\n   handoff where the amendment has one; preserve observed claims with acknowledged treatment.\n   Run `statusgen withdrawal-audit` over the candidate board and an open-PR snapshot before\n   rollout, then reconcile affected implementation and evidence PRs with their owners.\n   Do not auto-close them. Upgrade readers before authoring v3; see `docs\u002Fwithdrawal-rollout.md`.\n2. **Critical path** — an ASCII chain showing the blocking order, with a one-line \"smallest unblocking\n   move\" and a warning if any tempting-but-wrong first step is a dead end.\n3. **Dependency waves** — `Wave 0: [..]  Wave 1: [..]←0  …`, plus the one-line critical path\n   (`0→8→1→2→5`).\n4. **Delivery milestones when applicable** — link the canonical records and show outcome,\n   contributor wave span, remaining work, acceptance outstanding and prerequisite holds\n   per the shared procedure. Keep generated projections in their own supported region;\n   do not add milestone rows to the Briefs table or hand-edit generated cells. Until\n   supported, put the proposed mapping in the scoping document and link it from authored prose.\n5. (Optional) \"Before starting\" prereqs and \"Shared conventions\" the briefs inherit.\n\n## When a spec is approved: file the authoring follow-on in the same motion\n\nA spec or scoping doc reaching `approved` is not yet work — nothing pulls it forward until\nbriefs exist, so an approved-but-unrouted spec is exactly the thing that gets forgotten.\nClose that edge in the landing motion: whoever lands the PR that marks a spec `approved`\nfiles, in the same motion, ONE work-ready issue on your tracker titled **\"Author briefs for\n&lt;spec path&gt; into &lt;destination stream&gt; (strong tier)\"**, naming this\n`assay:author-brief` procedure and the **strong tier** requirement in the body so it rides\nyour existing issue → work-dispatch lane to a strong-tier author. \"Same motion\" means the\nlanding PR is not complete without it: file the issue immediately after the merge, or\nreference it from the PR body before merging. Landing those briefs later flips the spec's\nstatus to `routed` in the citing PR. (An automated emitter may back this floor as a second\nlayer; the floor does not wait on one.)\n\n## Working method\n\n1. **Reconcile outcomes and inventory the work** — preserve the source's milestone promises;\n   map criterion coverage to existing briefs before adding new units. Follow the shared\n   procedure for milestone dependencies, versions and reader support. Each new unit becomes\n   a brief; keep them roughly equal-\n   sized (split anything too big — e.g. one brief doing two distinct roles becomes two).\n2. **Map dependencies** — for each unit, what must exist first (`Depends on`) and what it enables\n   (`Unblocks`). This gives the waves (Wave 0 = no deps; Wave N = depends only on \u003CN).\n3. **Derive + verify the critical path** — for each selected milestone, include its contributing\n   work, actual execution dependencies and final acceptance prerequisites. Validate mixed-graph\n   cycles and qualify paths with unknown root\u002Fevidence coverage. Verify the head per the\n   judgment-call section above; a wave summary alone is not readiness.\n4. **Write the README first** (table + path + waves), then the briefs. Cross-link both directions.\n   Minimal skeleton, one `##` heading per section:\n\n   ```markdown\n   ## Briefs\n   ## Critical path\n   ## Dependency waves\n   ```\n5. **Convert relative dates to absolute** (e.g. \"by next week\" → the actual date) — these docs\n   outlive the conversation.\n6. **Author the facts and the edges, not the cell.** Status is DERIVED, not chosen: write `why:`,\n   `depends`, `unblocks`, `gate`, `risk` accurately and let `statusgen` read the PR trailers,\n   witnesses and approvals into the Status \u002F Verified \u002F Reviewed cells.\n   A project not yet wired to `statusgen` keeps the honesty\n   rule in spirit by hand — \"⚠ Workaround\" \u002F \"Blocked by #X\" beats a green checkmark that isn't\n   true — but once generation is live, a hand-edited cell is a lint PROBLEM, not a courtesy.\n7. **Handoff to execution**: a brief is a scope-and-DoD contract, not a step-by-step plan — don't\n   write one here. When a brief (especially Effort: L) is picked up for execution, consider running\n   `superpowers:writing-plans` against that brief's Deliverables\u002FDoD to produce a bite-sized TDD plan\n   before touching code. That's a separate step, done at pickup time by whoever executes it — not\n   part of authoring the brief.\n8. **Link the authoring PR with `Authors:`, never `Brief:`.** The PR that writes brief files\n   carries exactly one link line `Authors: \u003Cstream>\u002F\u003CNN>[, \u003Cstream>\u002F\u003CNN> …]` naming every brief it\n   adds (or `Issue: #\u003CN>` when the authoring answers an issue, or `Fixes: #\u003CN>` when merging the authoring PR should close it — never for an issue carrying `needs-decision` or `human-decided`, or any decision-gate issue; those link with `Issue: #\u003CN>`). `Brief: \u003Cstream>\u002F\u003CNN>` means the\n   PR DELIVERS that brief: the dispatcher's phantom check, the planner and the derived board all\n   read it that way, so an authoring PR carrying it makes the brief it just wrote look delivered\n   the moment it merges, and the brief is never dispatched. `deskpr create` refuses a `Brief:`\n   line on a branch that only authors that brief and names the `Authors:` line to use, and refuses\n   an `Authors:` line unless the branch adds every listed brief's file and touches nothing but\n   board READMEs, brief files and changelog fragments. A PR that authors a brief AND delivers\n   work in the same change keeps `Brief:`.\n\n## Conventions to inherit\n\n- Briefs are reference docs, not narration — terse, imperative, link to the authoritative source\n  rather than duplicating it.\n- When a brief's work lands, fill its Evidence section (one row per Verify item, filled by someone who\n  did NOT implement) and update the README status row + critical path in the same change, so the plan\n  never drifts from reality.\n- **Docs regeneration**: if this project has auto-generated docs or a docs-site that must be\n  regenerated when the public-facing architecture, API surface, user flow, or operational tooling\n  changes, every such brief MUST carry a docs-regeneration item in its DoD naming the specific\n  pages\u002Fareas to update. State the project's concrete regen command in the project-level skill (or\n  the repo's instructions file).\n- **Changelog fragment**: if the target repo enforces a per-PR fragment (a `changelog\u002F` directory\n  carrying `changelog\u002FREADME.md`), LIST `changelog\u002F\u003Cslug>.md` in the brief's `files:` line — `files:`\n  is the implementer's contract, and a path absent from it will not get written. The fragment is\n  human-legible highlight bullets, not a restatement of the Task; a genuinely non-notable brief says so\n  in its `why:` and leaves the `changelog:skip` waiver to the desk or a human. Do NOT add a Verify row\n  for it — the repo's CI changelog check IS that row.\n- **Short identifiers**: keep identifiers a brief asks for (test function names especially) under\n  32 characters, and in a PR body describe a long identifier rather than quote it: the desk secret\n  scan reads any 32+ character alphanumeric run as a possible secret.\n- If executing a brief surfaces a NEW non-obvious gotcha, fold it into the repo's instructions file\n  (CLAUDE.md \u002F AGENTS.md \u002F etc.) so the next person doesn't rediscover it.\n",{"name":5,"description":7},"main","7b56c62e2c2dc6a597a6b41ccdd43c72c70fe0ca","plugins\u002Fassay\u002Fskills\u002Fauthor-brief",5,"https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay","https:\u002F\u002Fgithub.com\u002Fmedici-finance","2026-11-15 00:52:35",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add medici-finance\u002Fassay --skill author-brief",{"tool":37,"command":38},"bunx","bunx skills add medici-finance\u002Fassay --skill author-brief",{"tool":40,"command":41},"pnpm","pnpm dlx skills add medici-finance\u002Fassay --skill author-brief","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Fplugins\u002Fassay\u002Fskills\u002Fauthor-brief\u002FSKILL.md",true,[45],{"path":46,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":49,"sourceHtmlUrl":50},"component.yaml","REFERENCE","参考文档",572,"https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Fplugins\u002Fassay\u002Fskills\u002Fauthor-brief\u002Fcomponent.yaml",[52,59,65,72,78,85],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":19},"01M7A2KDJHD3C240974C8CHEK0","Bad--Name","A valid description.","https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Ftools\u002Fskillslint\u002Ftestdata\u002Fconformance\u002Fname-pattern\u002FBad--Name\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":19},"01M7A2KCG3W4TB4ZTKK9NC3TX5","not-the-desk","https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Ftools\u002Fskillslint\u002Ftestdata\u002Fconformance\u002Fname-mismatch\u002Fthe-desk\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M7A2K70T9HZHH43ARFZP6P0A","skill","abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde abcde","https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Ftools\u002Fskillslint\u002Ftestdata\u002Fconformance\u002Fdesc-1025\u002Fskill\u002FSKILL.md",[],[],{"skillNo":73,"name":67,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7A2JYP6F7FCZK539GA01MFQ","字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字字 字字字字","https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Ftools\u002Fskillslint\u002Ftestdata\u002Fconformance\u002Fdesc-1024-multibyte\u002Fskill\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7A2JSBQENV3BBQZ3FQNSJ5E","skill-9","This is a filler sentence used only to push the summed bundle description character budget over its threshold for the fixture. This is a filler sentence used only to push the summed bundle description character budget over its threshold for the fixture. This is a filler sentence used only to push the summed bundle description character budget over its threshold for the fixture. This is a filler sentence used only to push the summed bundle description character budget over its threshold for the fixture. This is a filler sentence used only to push the summed bundle description character budget over its threshold for the fixture. This is a filler sentence used only to push the summed bundle description character budget over its threshold for the fixture. This is a filler sentence used only to push the summed bundle description character budget over its threshold for the fixture. This is a f","https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Ftools\u002Fskillslint\u002Ftestdata\u002Fconformance\u002Fbudget-over\u002Fskill-9\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M7A2JKR0EGK44TQVGZ1EWCW4","skill-8","https:\u002F\u002Fgithub.com\u002Fmedici-finance\u002Fassay\u002Fblob\u002Fmain\u002Ftools\u002Fskillslint\u002Ftestdata\u002Fconformance\u002Fbudget-over\u002Fskill-8\u002FSKILL.md",[],[]]