[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EXK8M9BERB6024TNJRGFC0":3,"skill-same:01M7EXK8M9BERB6024TNJRGFC0":49},{"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":29,"commitSha":30,"skillDirectory":31,"forksCount":14,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":6,"ownerDescription":34,"updatedAt":35,"install":36,"resources":48},"01M7EXK8M9BERB6024TNJRGFC0","orch-code-anth",null,"Plexar orchestration skill \u002Forch-code-anth: v5.0.1 (self-contained · computed-diff · deterministic apply · lane discipline)","NovemberFalls\u002Fplexar-harness-oss","NovemberFalls","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F70782979?v=4","01M7EXEMTMCAQSKSJ6878BEERP","Leonard D Boord III","01M7EXEMV9Q5YT78A4HW17N891",0,"MIT","https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss\u002Fblob\u002Fmain\u002Fpackages\u002Fbundle\u002Fplexar\u002Fskills\u002Forch-code-anth\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: orch-code-anth\ndescription: \"Plexar orchestration skill \u002Forch-code-anth: v5.0.1 (self-contained · computed-diff · deterministic apply · lane discipline)\"\nuser-invocable: true\nmetadata:\n  source: \"~\u002F.claude\u002Fcommands\u002Forch-code-anth.md\"\n  source_sha256: \"ab011be305274c9e23ad266067c09f3ef0d2f8be04515f8de21da04cd24e5ebc\"\n  ported: \"2026-09-22\"\n---\n\n> **Running in Plexar Harness.** This skill was written for Claude Code; read its tool names through this map.\n> - `Agent` tool → the harness `subagent` tool (in-process child agent; `wait_agent` \u002F `list_agents` to collect results).\n> - Lane aliases `haiku` \u002F `sonnet` \u002F `opus` → there is one primary rig model (`Qwen\u002FQwen3.6-35B-A3B`; escalation `prism-ml\u002FTernary-Bonsai-2-27B`); express a lane as effort instead: MUNDANE → `off`, WORKHORSE and CRITICAL → `high`.\n> - `Bash` \u002F `PowerShell` → `bash` \u002F `pwsh`; `Read` \u002F `Write` \u002F `Edit` → `read` \u002F `write` \u002F `edit`.\n> - `plexar_agents` (ledger, cards) is Python in `\u003Cworkspace>\\plexar-framework`; call it through `pwsh`\u002F`bash` with that folder on `PYTHONPATH`.\n> - Anything this skill names that the harness lacks (e.g. `ToolSearch`, `Monitor`) is unavailable here; do that step by hand or skip it and say so.\n\n# \u002Forch-code-anth — v5.0.1 (self-contained · computed-diff · deterministic apply · lane discipline)\n\n> ### §10 FIRES AT FIVE POINTS IN THE FLOW BELOW — NOT AT THE END WHERE IT IS WRITTEN\n>\n> `run_start` at **inventory** · `plan` at **plan freeze** · `dispatch` at **every spawn AND\n> every SOLO node** · `outcome` when **each report lands** · `run` at **landing**.\n>\n> §10 is mandatory and it is the last section in this file purely for readability. **A run\n> that reaches the inventory and writes nothing has already lost data that cannot be\n> recovered afterwards** — the 2026-09-21 PDF-builder run did exactly that, because §1 said\n> nothing about logging and §10 sat 420 lines later. **Read §10 before acting on §1.**\n\n\n**Patch class (2026-09-06).** v5.0.1 adds NO mechanism to the crowned v5.0. It repairs four\ndefects in the TEXT: (a) the skill was not self-contained — it referred the reader to\n\"v4.1 §1\", a document the reader does not have; (b) §7 and §8 gave contradictory commit\nrules; (c) the applier was named by a filesystem path that may not exist on the machine\nrunning this; (d) it carried machine-local absolute paths. v4.1 §1–§8 are now present here\nin full. The v5.0 mechanism is §0, §4.9, §4.9a — unchanged.\n\nCount the work, print a SOLO\u002FSWARM verdict, obey it. Split the work into DIFFABLE\n(expressible as verbatim text swaps) and GENERATIVE (everything else). Compute the\nDIFFABLE changes yourself and let a deterministic applier land them — no worker model in\nthe mechanical path. Gate on a deterministic check, never on a model's opinion.\n\n**Operating principle.** The orchestrator is the brain. When it already understands the\nexact change, it pushes that change out as data rather than delegating the *understanding*\nto a worker. Application of a verbatim diff is deterministic — it does not need a model.\n\n**Self-containment rule.** Everything this skill requires you to do is written HERE. If you\nfind yourself reasoning \"the rule is in another document I don't have, so I'll infer it,\"\nthat is a defect in this file — report it. Do not invent the missing rule.\n\nYou are the ORCHESTRATOR: Opus-class judgment at the top, the cheapest model that clears\nthe gate underneath. Your spend follows your typing. Objective order when they conflict:\n**correctness → premium-token displacement → wall-clock → raw dollars.**\n\n**Orchestrator effort.** Two configurations, pick by the job. `opus \u002F high` when ONE PASS\nMUST LAND with no retry loop. `opus \u002F low` — the DEFAULT — when a deterministic gate plus\nthe §4.8 escalation ladder are watching, which is this skill's own operating mode: at the\nbottom of the ladder the orchestrator simply executes the computed plan without paying for\ndeliberation the skill already did. `medium` is price-dominated — never park in the middle.\n`xhigh` over-thinks. Workers: the card is the quality mechanism; the model only buys\nwall-clock.\n\n*Measurements, corrections and known limits are deliberately NOT in this file. If the bench\nrepo is present they are in `team\u002Fbench\u002Fresults\u002F` (`V4_FINDINGS.md`, `ROUND3_FINDINGS.md`,\n`V5_LIVE_FINDINGS.md`) and Agents-Public `FINDINGS.md` (its own sections 4.7 and 4.7a —\nnot this file's §4.7); published at\n\u003Chttps:\u002F\u002Fboord-its.com\u002Fskills>. Those pointers are evidence, not instructions — this file\nis complete without them.*\n\n---\n\n## §0 · The apply-tier gate (decide after §1, before choosing SOLO\u002FSWARM)\n\nAfter the §1 inventory, ask one added question of the work: **is a node's change\nEXPRESSIBLE as localized, verbatim text swaps against files that already exist?**\n\n- **DIFFABLE** — logging\u002FAPI migrations, signature threading, rename sweeps, config edits,\n  mechanical refactors: the change is N independent site-edits with a computable\n  before\u002Fafter. These are the apply-tier's target.\n- **GENERATIVE** — new files, structural rewrites, algorithm design, anything whose output\n  isn't a swap against existing text. A model implements these; the apply-tier does not\n  apply.\n\nPrint: `APPLY-TIER: \u003CDIFFABLE n=\u003Csites> | GENERATIVE | MIXED c=\u003Cdiffable>\u002F\u003Cgenerative>>`\n\nMIXED is normal: the CRITICAL\u002FGENERATIVE core (e.g. a shared contract rewrite) is\nimplemented by an opus worker per §4.4; the DIFFABLE fan-out (the dozens of call-site\nmigrations that depend on it) routes to the apply-tier. This is the same split §4.2 already\ndraws between the pinned-contract core and its downstream sites — the apply-tier just makes\nthe downstream deterministic instead of a worker swarm.\n\n## §1 · Inventory (mandatory, before choosing anything)\n**→ FIRST, before anything below: write the `run_start` row (§10).**\n\n\nRead the task and the spec\u002Fsource. COUNT the real work — do not estimate from the prompt\nalone:\n\n- **sites** — individual edit\u002Fcreation points (grep the call sites; list the stubs)\n- **files** — files you must create or modify\n- **read volume** — bytes of code\u002Fspec that must actually be read (`wc -c`, ÷4 ≈ tokens)\n- **nodes** — discrete deliverables, labeled by RISK (not size):\n  - **CRITICAL** — shared contracts others import · auth\u002Fsecurity · money math ·\n    concurrency · migrations · anything DESTRUCTIVE (§6)\n  - **WORKHORSE** — contained implementation against a clear spec · trap-dense edits\n  - **MUNDANE** — mechanical, fully specified, judgment-light\n\n## §2 · Scale gate (computed — print the verdict, then obey it)\n\nPrint one line: `GATE: SOLO|SWARM — sites=\u003Cn> files=\u003Cn> read≈\u003Cn>K nodes=\u003Cn> mix=\u003CC\u002FW\u002FM>`\n\n**SWARM** if ANY of:\n\n- sites ≥ 25, or files to create\u002Fmodify ≥ 12\n- nodes ≥ 12 AND ≥ 40% of them are MUNDANE+WORKHORSE (displacement value exists)\n- read volume ≥ 150K tokens, or projected work ≥ 60% of the session's turn budget\n- the human flagged premium-token pressure\n\n**SOLO** otherwise — that is the measured optimum below these lines, not laziness.\n\n**Tiebreak:** when a count sits within one honest re-count of a line (e.g. 9 vs 10\ndeliverables), take SOLO — below the wall the cheaper error is under-routing.\n\n**Mid-run re-gate (the insurance the monolith lacks):** while SOLO, if you reach half the\nturn budget with less than half the checklist done, or discovered sites exceed 1.5× your\ninventory, STOP — write the plan (§4.1) for the REMAINDER and switch to SWARM.\n\nThe apply-tier does not change WHEN to swarm — it changes HOW the DIFFABLE portion of the\nwork is executed.\n\n## §3 · SOLO path\n\nDo the work in-session with the discipline that survives the wall: write the full checklist\nfirst (every site, `file:line`), execute in file-order batches, tick each site off, re-grep\nafter each batch to catch missed sites, then run the §5 gate. Interactively, you may\ninstead hand the whole job to ONE domain specialist when the main context is precious —\nbrief per §4.5, same gate.\n\nIf the SOLO job is DIFFABLE and large, use the apply-tier (§4.9) on yourself: compute the\nblocks, apply deterministically, gate. That is where the apply-tier's speed and cost win\nshows up even without a swarm.\n\n## §4 · SWARM path\n\n### §4.1 · Plan-as-data\n\nCreate a scratch dir OUTSIDE the repo (`mktemp -d`, or `%TEMP%\\orch-\u003Cname>`). Write\n`plan.json` there before any implementation:\n\n```json\n{\"mode\":\"swarm\",\"gate\":\"GATE line verbatim\",\n \"workers\":[{\"id\":\"W1\",\"nodes\":[\"N01\"],\"lane\":\"opus\",\"files\":[\"lib\u002Flog.js\"],\n             \"deps\":[],\"spec\":\"SPEC.md 2.1-2.4\",\"destructive\":false,\"apply_tier\":false}],\n \"batches\":[[\"W1\",\"W2\"],[\"W3\"]],\n \"gate_cmd\":\"\u003Cthe 4.7 script>\"}\n```\n\nNOTHING non-product is ever written inside the repo — no plan, no scratch, no gate script,\nno self-check artifacts. The repo receives product code only.\n\n**Output diet (your tokens are the expensive ones):** plan.json ≤ ~40 lines; each worker\ncard ≤ ~12 lines beyond the shared preamble; reconciliation ≤ ~30 lines; gate script\nfocused (contracts + named traps + spot-sites + one integration path, ≤ ~150 lines) — the\nper-worker CHECKs already covered the breadth. Do not restate preamble content in cards; do\nnot narrate between tool calls.\n\nPresent the plan (worker table + lanes + batches). If a human is present, pause for\napproval; in auto-approve\u002Fheadless mode print it and proceed (`PLAN-APPROVED: AUTO`).\n\n### §4.2 · Group into workers, collapse tiers\n\n- **One worker per cohesive cluster** — a subsystem\u002Ffile family of ~3–8 sites or 1–3\n  modules — NOT one worker per graded requirement. Target 6–12 workers for a 15–70-site\n  job. Never two writers of the same file in the same batch.\n- **Group WITHIN a lane, never across lanes.** Folding a MUNDANE node into a sonnet or opus\n  worker silently up-lanes it — that is where the haiku lane went in v4.0. A worker whose\n  nodes are all MUNDANE runs on `haiku`; promoting it to sonnet requires a trap you can NAME\n  in that worker's own files, named in plan.json.\n- **Contract-pinned tier collapse:** a dependency whose interface is pinned by a normative\n  spec (exact signatures, values, worked examples) is NOT a batch edge — downstream workers\n  code against the pinned contract while the upstream is built in parallel. Only\n  physically-shared prerequisites (foundation modules everyone imports, with no pinned spec)\n  create a second batch. Most swarms are 1–2 batches; minimize the critical path — two\n  correct plans can differ 2x in wall-clock.\n\n### §4.3 · THE MANDATE (hard rule on this path — the three excuses are pre-refuted)\n\n- You do NOT implement GENERATIVE work. Every such worker is dispatched via the `Agent`\n  tool. \"It's quicker to just do this one myself\" is the defection the arena caught 28\u002F28\n  times.\n- \"Spec fidelity\" is preserved by the brief plus the worker reading the spec FROM DISK — not\n  by hoarding the work. \"It fits in one context\" is answered by the gate, which already said\n  SWARM. \"Per-node briefing cost\" is answered by grouping (§4.2) and the shared preamble\n  (§4.5).\n- Dispatch every ready worker of a batch as parallel `Agent` calls in ONE message — all the\n  tool calls in the SAME assistant turn (≤ ~10 per message; split larger batches).\n  **Spawning one worker, awaiting its result, then spawning the next is a mandate violation,\n  not prudence** — v4.0 did exactly that 10\u002F10 times and turned the swarm into a relay race.\n  There is nothing to check between same-batch spawns: landing checks (§4.6) run after the\n  WHOLE batch has returned, never between spawns.\n- Pass `model` EXPLICITLY on every spawn — never inherit. Put the worker id and node ids in\n  `description` (e.g. `\"W3 N04-N07 mundane sweep\"`) so routing is auditable.\n- Your own turns are reserved for: inventory, the plan, briefs, computing apply-tier blocks,\n  reading reports, landing checks, the gate, escalations, reconciliation.\n- **The apply-tier is not an exception to the mandate.** It is deterministic execution of\n  work you have FULLY computed as verbatim data, which is categorically different from\n  quietly doing a worker's judgment job. If you are making judgment calls while \"applying,\"\n  you are implementing — dispatch it.\n\n### §4.4 · Lanes (rubric, decided once in the plan)\n\n| capability class | tier | model alias | goes there |\n|---|---|---|---|\n| CHEAP | MUNDANE | `haiku` | mechanical and fully-ruled, WITH the exact rules and trap warnings in the card |\n| WORKHORSE | WORKHORSE | `sonnet` | contained implementation; also trap-DENSE mundane clusters |\n| CRITICAL | CRITICAL | `opus` | shared unpinned contracts, auth, money, concurrency, destructive |\n| DETERMINISTIC | DIFFABLE fan-out | *(none)* | orchestrator computes blocks → §4.9 applier. No model. |\n\nThe capability class is the architecture; the alias is today's mapping. When you change\nwhich model wears a jersey you have changed the mapping — not the algorithm — and the\nnumbers behind the mapping do not carry.\n\n**Demotion rule:** a CRITICAL-labeled node whose behavior is fully pinned by the spec\n(signatures + values + worked examples) MAY run on `sonnet` — but NEVER auth, money,\nconcurrency, or destructive nodes. **Promotion rule:** a mundane cluster dense with\nanti-mechanical traps runs on `sonnet`. `subagent_type: general-purpose` always works; a\nmatching domain specialist is fine too — the explicit `model` still governs.\n\n**Pinned generation.** Spawns use tier aliases (the `Agent` tool accepts only\nhaiku|sonnet|opus|fable), but every number in the evidence files describes ONE pinned\ngeneration — **G1**: haiku=`claude-haiku-4-5-20251001`, sonnet=`claude-sonnet-5`,\nopus=`claude-opus-4-8`, seat experiments=`claude-fable-5`. When a provider release re-points\nany alias, the numbers do NOT carry: run the release-qualification protocol (champion cells\nat k=2, then the min-max dials) BEFORE adopting or trusting. A release is a fresh min-max\nopportunity, never a silent upgrade. **As of 2026-09-06 the `opus` alias no longer resolves\nto G1. Every figure in the evidence files is therefore UNQUALIFIED until that protocol is\nre-run.**\n\n### §4.5 · Briefs (cache-shaped; pointers, not pastes)\n\nAssemble ONE shared preamble, byte-identical across every spawn of the run (identical\nprefixes are prompt-cache hits — the swarm pays for it once):\n\n```\nRepo root = your working directory. SPEC.md is the binding contract — READ the\nsections named in your card FROM DISK; they are normative, never paraphrased here.\nGlobal constraints: \u003Cthe task's own rules: stdlib-only, no new files outside X, ...>\nSelf-check before reporting, with all scratch files OUTSIDE the repo (system temp).\nReport EXACTLY:\nSTATUS: DONE | BLOCKED(\u003Cquestion>)\nFILES: \u003Ctouched>\nCHECK: \u003Ccommand> -> \u003Cresult, 1 line>\nNOTES: \u003C=3 lines\nIf any instruction cannot be applied exactly as written: STATUS: BLOCKED. Never guess.\n```\n\nThen a per-worker card: worker id · OWNED files (exact list — touch nothing else) · spec\nsections BY REFERENCE · the byte-exact rules and trap warnings that bite (these are the only\npasted lines — error templates, orderings, \"must NOT change\" items) · upstream state\n(\"fieldops\u002Fcontracts.py is FINAL on disk — import it, never modify\") · done-looks-like · the\nself-check command to run.\n\n### §4.6 · Landing protocol (reports are claims)\n\nAfter the WHOLE batch returns (all spawns of the dispatch message — never between spawns):\nrun `plexar_agents.scope.sweep()` (it sees Bash writes a hook cannot) — any changed file outside the\nunion of dispatched OWNED lists is reverted and the owner re-briefed. Spot-check each worker\nwith its own CHECK command or a 1-line import\u002Frun probe. A BLOCKED report gets an answer\n(from spec\u002Fplan), then a fresh spawn, same lane. Never accept a green self-report as the\ngate — in the arena a correct diff arrived under a wrong report, and only the script told\nthe truth.\n\n### §4.7 · The gate (orchestrator-owned, deterministic)\n\nWrite your OWN gate script in the scratch dir — worked examples from the spec made\nexecutable, one end-to-end integration path, the anti-regression traps (\"must not change\"\nitems verified verbatim), and the scope sweep. **Exit code decides.** Non-zero blocks: fix\nvia §4.8 and re-run. LLM review is optional and never substitutes.\n\n**The gate is not answerable to the implementation.** A failing gate means the code is wrong\nuntil you produce independent evidence that the CHECK ITSELF is malformed — the\nimplementation disagreeing with the gate is never that evidence. If the gate is genuinely\nbroken, or cannot observe the behavior the spec promises, stop and say so; do not weaken it\nand declare a pass.\n\n### §4.8 · Escalation ladder (evidence travels, work doesn't restart)\n\nPer failing worker: (1) fresh spawn, SAME lane, with the gate failure plus the current diff\nof its owned files pasted in; (2) one lane up (haiku→sonnet→opus), same evidence; (3)\n`LADDER_EXHAUSTED` — you repair the minimal failing diff in-session and log it in\nreconciliation. The cap is counted by `plexar_agents.ladder.attempt()`, which raises on the fourth.\n\n### §4.9 · The apply-tier (for the DIFFABLE portion only)\n\nWhen the plan marks a node or cluster DIFFABLE:\n\n1. **Compute, don't delegate.** The orchestrator reads the target files and the binding spec\n   and emits one **SEARCH\u002FREPLACE block per site**, in exactly this format:\n\n   ```\n   FILE: \u003Cpath from repo root>\n   \u003C\u003C\u003C\u003C\u003C\u003C\u003C SEARCH\n   \u003Cexact current line(s), copied VERBATIM from the file — byte-for-byte, same indent;\n    include enough surrounding lines to be UNIQUE if a bare line repeats>\n   =======\n   \u003Creplacement line(s) per the spec rules>\n   >>>>>>> REPLACE\n   ```\n\n   The SEARCH text MUST be verbatim from disk — that is the entire fidelity mechanism.\n   Approximate SEARCH is a defect worse than no diff: measured, a worker handed a lossy\n   \"exact\" patch scored BELOW the same worker reading the spec. Emit this as plan data to\n   the scratch dir (`plan.blocks`), never narrated inline.\n\n2. **Apply deterministically — no worker model.** Run an applier meeting this contract:\n\n   - A block applies **iff** its SEARCH text is found **exactly once** in the current\n     in-memory state of the file. Zero matches → `nomatch`. Two or more → `nonunique`.\n   - Non-matching blocks are **NEVER** force-applied, fuzzed, or eyeballed into place.\n   - Blocks for one file apply sequentially against the evolving in-memory text (a later\n     block sees earlier blocks' results); the file is written once, at the end.\n   - Line endings are preserved. Try an exact byte match first; count an EOL-normalized\n     match separately so CRLF\u002FLF drift stays visible rather than hidden.\n   - Report `applied \u002F nomatch \u002F nonunique` counts and per-block residue.\n   - Support an `--atomic` mode that refuses partial application (§4.9a needs it).\n\n   A clean run is `applied == blocks, nomatch=0, nonunique=0`. This step costs ~0 tokens and\n   ~0 seconds and is 100% reproducible.\n\n   **Implementation.** If `tools\u002Fclean\u002Fapply_blocks.py` or\n   `~\u002F.claude\u002Ftools\u002Fclean\u002Fapply_blocks.py` exists, USE IT — do not re-implement. If neither\n   exists, write a short stdlib script to the contract above in the SCRATCH dir, never in\n   the repo. The contract is normative; the file path is a convenience.\n\n3. **Residue → escalate, don't guess.** Any `nomatch` (your SEARCH wasn't verbatim) or\n   `nonunique` (insufficient context) block does NOT get force-applied. Either re-emit that\n   block yourself with more surrounding context, or hand the single block plus the file to\n   one cheap worker: \"apply this one edit.\" Deterministic-first; model only for the residue.\n\n4. **Gate as always (§4.7).** The deterministic apply changes nothing about the gate — the\n   script, not the block counts, decides done. A verbatim-clean apply is necessary, not\n   sufficient: you can still compute a WRONG replacement (a planning miss), which only the\n   gate catches.\n\n### §4.9a · Fleet reuse (the scale multiplier)\n\nBecause the SEARCH\u002FREPLACE set is data and the apply is deterministic, the SAME computed\ndiff applies to N repositories or instances at ~0 marginal cost. Compute once; apply\neverywhere. The compute is spent on understanding ONCE, not re-litigated per target.\n\n**A fleet needs its own policy, because per-target success does not imply fleet success.**\n93 of 100 clean with 7 residue is not \"93% done\" — it is a fleet running two versions.\nDeclare the policy in plan.json before applying:\n\n- `FLEET_POLICY: ATOMIC` — every target must apply clean or NONE are kept; any residue rolls\n  the whole fleet back. **This is the default, and it is mandatory** for anything touching a\n  shared contract, a wire format, auth, money, or a migration.\n- `FLEET_POLICY: PARTIAL_ALLOWED` — clean targets are kept and residue targets are\n  QUARANTINED: named explicitly in reconciliation, with the resume condition. Permitted only\n  for independent, non-contract edits.\n\nPrint `FLEET: \u003Cpolicy> targets=\u003Cn> clean=\u003Cn> residue=\u003Cn> quarantined=\u003Clist|NONE>`. A fleet\nrun with unreported residue is a failed run.\n\n## §5 · SOLO gate\n\nSame as §4.7 — the script, not your eyes, decides done. Run it before declaring, and the\nsame \"the gate is not answerable to the implementation\" rule applies.\n\n## §6 · Destructive interlock\n\nDESTRUCTIVE = schema\u002Fdata migrations, data deletion or movement, git history rewrites or\nforce-push, deploy\u002Finfra mutation, secret or auth-provider changes — anything irreversible\noutside the working tree. Rules, regardless of path:\n\n- never routed below `opus`; the worker returns PLAN plus exact commands plus diff plus\n  rollback — it does not execute;\n- you verify the rollback actually restores state before presenting;\n- a live human approves execution. In auto-approve\u002Fheadless mode the irreversible step is\n  NOT executed: mark it `DEFERRED-DESTRUCTIVE` with the ready-to-run plan.\n\nA computed diff touching auth, money, concurrency or migrations is STILL destructive. The\napply-tier does not launder it: it is presented as a plan with rollback and never\nauto-applied.\n\n## §7 · Reconcile and hand off\n\n- **Requirement matrix:** every spec item → IMPLEMENTED \u002F PARTIAL \u002F DEFERRED \u002F SCOPE-CREEP,\n  with file refs.\n- **Routing table:** planned vs ACTUAL spawns (worker · lane · model · attempts) plus any\n  `LADDER_EXHAUSTED` or `DEFERRED-DESTRUCTIVE` events. Planned not equal to actual must be\n  explained. DIFFABLE nodes carry `blocks applied \u002F nomatch \u002F nonunique` and any residue\n  escalations.\n- **Gate evidence:** the script's final output tail.\n\n### §7.1 · Commit policy — ONE rule, stated once, nowhere else\n\nThe boundary is **publication**, not branch name:\n\n| action | approval |\n|---|---|\n| local candidate commit (not pushed) | **none required**, unless §6 applies |\n| push · PR · merge to a shared branch · release · deploy | **human approval, always** |\n| auto-approve \u002F headless mode | **no commits at all** — leave the tree uncommitted |\n\n**Avoid towering commits.** Each node or DIFFABLE cluster lands as its own local candidate\ncommit — a rollback and attribution boundary, not publication — tagged\n`[UNREVIEWED][\u003Cnode>][\u003CCLASS>] \u003Csummary>`. Do NOT accumulate a whole fan-out into one giant\nworking-tree diff; the plan's per-node data already gives you the boundaries, so commit along\nthem.\n\nIf any other section of this file appears to state a different commit rule, this table wins\nand the other section is a defect — report it.\n\n## §8 · Hygiene and optional add-ons\n\nCleanliness the core already enforces by construction: nothing non-product ever enters the\nrepo; every landing runs the scope sweep and reverts strays; SCOPE-CREEP is a named\nreconciliation state. Beyond that, after the gate passes and before handoff, you MAY add\n(interactively ask; pre-authorized runs note it in reconciliation):\n\n- **Hygiene sweep (deterministic-first):** run the repo's own tooling when present (ruff \u002F\n  eslint \u002F knip \u002F vulture \u002F tsc-unused ...) on the changed surface; what tooling can't see\n  (orphaned files, dead branches, stale docs touched by this change) goes to ONE cheap\n  worker scoped to THE DIFF ONLY — never a whole-repo crusade mid-task.\n- **LLM review:** optional second read for hygiene, architecture, or requirement-fit — never\n  the correctness gate (LLM reviewers hallucinate exactly where the deterministic gate\n  doesn't).\n- **Security — NOT optional:** any diff touching auth, sessions, input handling, or secrets\n  gets a security read before handoff.\n\nCommit behavior in this section is governed by §7.1. This section states no commit rule.\n\n## §9 · The clean phase — WITHDRAWN 2026-08-05, pending real measurement\n\n**This skill has no clean phase. Do not run one as part of a run.**\n\n§9 shipped mandatory, was demoted to optional on 2026-07-31, and is now out entirely. The\none measurement it has is not evidence it works: on `arena_cleanup` at k=3 it cost +30% and\n+34% wall for zero correctness gain — but every arm hit the 16\u002F16 ceiling *including the arm\nwith no clean phase at all*, because that fixture's prompt asks for cleanup, so a plain arm\njust does it by hand. The fixture left no headroom. So the honest state is **unmeasured**,\nnot measured-null, and an unmeasured phase does not ride along in the champion skill. If it\nearns its keep on a fixture with headroom, it comes back with numbers attached.\n\nCleanup is still available — as a deliberate, separately-invoked `\u002Forch-clean` when you have\na REASON (an untraced rename or signature migration, a run that removed or commented out\ncode, a repo with known half-migrated state). That is a decision you make, not a tax every\nrun pays. When you do invoke it, its ordering rule stands: **gate → clean → re-gate**,\nbecause cleanup deletes code and the gate that ran before the deletion has verified nothing.\n\n## §10 · The dispatch log (MANDATORY — this is training data, not telemetry)\n\n**Block version — log it verbatim as `skill_version`:** `s10-v11-2026-09-21`\n\nBump that string whenever this block changes. `skill_version` is THIS block's version, not\nthe skill's title-line version — they are different facts and the first pilot conflated them\n(logged `v5.0.1`, which describes the skill and says nothing about the schema that wrote the\nrow). Log the skill's own version separately as `skill_title_version` if you want both.\n\n### Where things go\n\n```\nlog      \u003Cworkspace>\\plexar-telemetry\\runs\\dispatch-\u003CYYYY-MM-DD>.jsonl\ncards    \u003Cworkspace>\\plexar-telemetry\\artifacts\\\u003Crun_id>\\cards\\\u003Cworker>.txt\n```\n\n**Paths, MSYS rewriting and the scratch-dir rule are handled by `plexar_agents.ledger`.**\nIt writes native Windows paths and never asks you to remember one. *(Was 13 lines.)*\n\nBefore each spawn, write **the full prompt as sent** — shared preamble plus card body — to\n`artifacts\\\u003Crun_id>\\cards\\\u003Cworker>.txt`. Log `brief_path` and `brief_sha256`; do **not** try\nto inline the card into JSON.\n\nThis exists because `brief_text` is not honestly populatable: the orchestrator has no\nprogrammatic handle on a prompt it already sent, `Agent` returns an id and an `output_file`,\nand those transcript files are empty. The only route to verbatim would be retyping a 60-line\ncard into a JSON string per spawn — premium tokens duplicating text that exists, and exactly\nthe step a future orchestrator skips silently, producing the summary this schema forbids.\nWriting the file first makes verbatim **free**, and the hash makes it checkable.\n\n**Card writing, LF normalisation, hashing and `retry_mode` are handled by\n`plexar_agents.cards.write()`.** It hashes the bytes on disk, so the hash cannot\ndisagree with the file. A resumed worker passes `is_delta=True`. *(Was 13 lines.)*\n\n**`run_start`** — ONE row at INVENTORY, before planning. `run_id` · `skill` ·\n`skill_version` · `cwd` · `task_summary`. Exists so an empty `runs\\` directory is\nunambiguous: no row means no run, rather than \"possibly still planning\". Costs nothing.\n\n**`plan`** — ONE row at plan freeze: `path` (SOLO|SWARM) · `scale_gate_reason` ·\n`node_count` · `lanes_planned` · `efforts_planned` · `edges` · `max_depth` ·\n`interfaces_declared` · `interfaces_consumed` · `interfaces_absent_at_dispatch` ·\n`diffable_sites` · `generative_sites` · `apply_tier_nodes` · `gate_script_sha256` ·\n`preamble_sha256`.\n\n**`dispatch`** — at each spawn, **and once per node on the SOLO path too** (see below):\n`node_id` · `worker` · `phase` · `lane` · `effort` ·\n`lane_reason` · `brief_path` · `brief_sha256` · `brief_author_model` · `files_owned` ·\n`spec_sections` · `diffable` · `retry_mode` · `pred`.\n\n**`outcome`** — when the report lands: `status`\n(`DONE|BLOCKED|GATE_FAIL|LADDER_EXHAUSTED`) · `blocked_question` · `blocked_cause` ·\n`attempts` · `escalated_to` · `files_touched` · `scope_violation` · **`node_check_cmd`** ·\n**`node_check_exit`** · `wall_ms`.\n\n**`run`** — ONE row at landing, `node_id: \"RUN\"`: **`status`**\n(`COMPLETED|HALTED_FOR_DECISION|ABANDONED`) · `gate_exit` · `gate_failing_check` ·\n`gate_evidence` · **`gate_verified_against_defect`** ·\n**`interfaces_declared_uncalled_at_landing`** · `orchestrator_repairs` ·\n`orchestrator_repair_notes` · `wall_ms`.\n\n### `HALTED_FOR_DECISION` — a run that correctly refused to dispatch is still a run\n\nThe node statuses (`DONE|BLOCKED|GATE_FAIL|LADDER_EXHAUSTED`) describe work that was\ndispatched. **A run that stops at the plan stage to put a decision to the human dispatched\nnothing and is none of them** — so before v5 it was unrepresentable, and therefore unlogged.\n\nThat is not a gap, it is a **bias in one direction.** Those runs are exactly the cases where\nthe orchestrator judged the task not ready and asked first — the same judgement a\nbrief-quality classifier exists to make. **Train on data that excludes them and the model\nnever learns to recommend stopping, because it has never seen stopping rewarded.**\n\n* `COMPLETED` — the gate ran and the run landed.\n* `HALTED_FOR_DECISION` — stopped before dispatching to put a choice to the human. Write the\n  `plan` row with whatever the assessment produced, and record the question in\n  `gate_evidence`. **This is a legitimate and often correct outcome; log it as one.**\n* `ABANDONED` — stopped for any other reason. Say why in `gate_evidence`.\n\nFound 2026-09-21: an `\u002Forch-code-anth` run assessed a PDF-builder request, proposed a design,\nand stopped because the approach changed the data model. Best judgement in the transcript,\nand it wrote zero rows.\n\n### `gate_exit` is a RUN field. Nodes carry `node_check_exit`.\n\n§4.7 mandates ONE gate script for the run, so a per-node `gate_exit` is **structurally null\nforever** — the first pilot had it null on 6 of 7 outcome rows, which made the KPIs\nuncomputable as specified and forced them to be read off `status`, which `kpi.md` forbids.\n\nThe per-node signal already exists: §4.5 requires every worker report\n`CHECK: \u003Ccommand> -> \u003Cresult>`. Log that command and its exit as `node_check_cmd` \u002F\n`node_check_exit`. The run's gate exit goes on the `run` row, where the evidence actually is.\n\n### `gate_verified_against_defect` — three states, never two\n\n**When a node exists because a specific defect escaped, run the gate once against the\nREINTRODUCED defect before calling the run green.** A gate nobody has seen fail is an\nuntested instrument, and §4.7 stakes everything on the exit code meaning something. The first\npilot did this in two minutes and it is the only reason \"the suite observes the bug it names\"\nis a measurement rather than a claim.\n\n- `true` — defect reintroduced, gate failed, defect restored, suite green again.\n- `false` — not attempted. This is a gap, and it should look like one.\n- `\"not-reproducible-in-harness\"` — the defect genuinely cannot be reintroduced in a test\n  (needs a real WebView, a sub-frame race, a stale cookie). **A binary field would silently\n  convert \"I could not verify this\" into \"I did not bother\", which inverts its purpose.**\n\n### `blocked_cause` — and it must cover guidance that MISLEADS, not only guidance that BREAKS\n\nEvery BLOCK gets one of:\n\n- `brief` — the card was genuinely silent or wrong. **The only value a brief-quality\n  classifier should train on as a positive.**\n- `orchestrator-broken` — the orchestrator broke the node's target between batches. The\n  worker was right to refuse.\n- `orchestrator-misleading` — the guidance compiled, was accurate enough to act on, **and was\n  still materially wrong.** The first pilot's own unblocking delta placed a call in a class\n  whose dependency is null in production: it would have passed every test and been a no-op on\n  the real device. **A card that compiles and misleads is the more dangerous of the two** and\n  no test or gate will find it.\n\nWithout this field, every mid-run orchestrator repair depresses the exact number a\nbrief-quality classifier trains on, charging workers for the orchestrator's mistakes.\n\n### `interfaces_*` — typed, and checked at BOTH ends\n\nEach entry is `{kind, name, path}`, `kind` one of:\n`function` · `method` · `channel` · `http_route` · `env_var` · `payload_shape` · `cli_flag` ·\n`config_key`. The first pilot hit five kinds in one run, and because\n`interfaces_absent_at_dispatch` is set arithmetic over free text, it came out `[]` **only\nbecause the same spellings happened to be chosen at both ends.** Untyped, this field reports\nsuccess by luck.\n\nTwo checks, neither needing a model:\n\n- **`interfaces_absent_at_dispatch`** (plan row) — consumed but not on disk and not produced\n  by an earlier tier. A failure prediction from pure set arithmetic.\n- **`interfaces_declared_uncalled_at_landing`** (run row) — one grep per declared interface\n  after the batch. Catches **built-but-never-wired**: the first pilot had an interface in\n  `interfaces_declared` AND `interfaces_consumed`, both ends built, and **zero call sites**,\n  because no consumer's card mentioned it. Shipped, that ack's deadline would have expired on\n  every SUCCESSFUL locked turn and reported failure on the exact case it was built to\n  confirm. **Worse than omitting it**, and it survives green worker reports and a green gate.\n\n**Consequent rule on §4.5:** an interface in `interfaces_declared` must appear in at least one\n**consumer's** card, not only its producer's. A promise nobody was told to call is not wired.\n\n### SOLO writes dispatch rows too, or its outcomes are untrainable\n\n**On the SOLO path the orchestrator IS the worker, and it still writes one `dispatch` row per\nnode** — `worker: \"self\"`, `lane: \"none\"`, `lane_reason` recording why SOLO was chosen for\nthis node, and `brief_path` pointing at the intent written to a card file exactly as a spawn\nwould.\n\n**`effort` IS THE EFFORT ALONE — never a model\u002Feffort pair.** On a `self`\u002F`none` node there is\nno worker model, and orchestrators have been filling `effort` with their OWN capability as\n`\"opus\u002Flow\"`. That information is worth keeping, but writing it there destroys the field:\n`\"opus\u002Flow\"` is not comparable to `\"low\"`, so those rows drop out of the one analysis `effort`\nexists for. Write `effort: \"low\"` and put the orchestrator's own model in\n**`orchestrator_model`**. Measured 2026-09-21: 5 rows were lost this way before the field was\nsplit.\n\nThe second pilot run was SOLO and produced **three `outcome` rows and zero `dispatch` rows**.\nThat leaves outcomes with **no corresponding input**, which breaks the (input, outcome) pairing\nthis entire log exists to produce: an outcome nobody can attribute to a brief teaches a\nclassifier nothing. **A SOLO run that logs only outcomes has logged nothing trainable.**\n\n### The field rules\n\n- **`ts`, JSON escaping and append-only writing are `plexar_agents.ledger`'s job.**\n  Every row carries an offset because the writer supplies it. *(Was 9 lines.)*\n\n- **`brief_path` + `brief_sha256`, never a summary.** The card is the input half of every\n  (input, outcome) pair. If you find yourself paraphrasing, the file step was skipped.\n- **`brief_author_model` is who WROTE the card, not who ran it.** Without it, a bad card and a\n  bad card-*writer* are indistinguishable and anything trained on this averages a confounder.\n- **`effort` is a separate axis from `lane`.** The target is `(brief, model, effort) ->\n  P(success)`; effort is the cheap dial that turns a BLOCKED into a DONE without moving tier.\n- **`lane_reason` is a forcing function, not bookkeeping.** Measured: writing \"contained\n  native change against a pinned name; no new concurrency beyond one cancellable deadline\"\n  demoted a node from opus to sonnet *while it was being written*. Write the clause that\n  decided it, not the tier.\n- **`blocked_question` verbatim** from `STATUS: BLOCKED(\u003Cquestion>)` — the label for which\n  field the brief omitted, in the worker's own words.\n- **`files_owned` vs `files_touched`** — the landing protocol already computes this diff to\n  revert strays; recording it turns enforcement into a free label.\n- **`orchestrator_repairs` is mandatory, including zero.** If the log cannot see the\n  orchestrator's own defects it will systematically attribute them to workers. The first\n  pilot had three, and **all three would have passed both the worker reports and a green\n  gate.**\n- **`node_check_exit: null` is not `0`.** A check that did not run is not a check that passed.\n- **`pred` is expected-null until a classifier exists.** Nulls here are the designed state at\n  row 0, not missing data. Absent predictions are `null`, **never `0.0`** — a missing opinion\n  and a confident zero are opposite claims. The rubric carries no `p_success` at all: it is a\n  decision table, and inventing one would make the comparison dishonest.\n\n### Rig-dispatched workers carry the join on the WIRE (contract §10h)\n\nEverything above assumes the worker was spawned through the `Agent` tool, where a\n`PreToolUse` hook can capture the dispatch deterministically. **`\u002Forch-code-local`\ndispatches over HTTP to the rig instead, so no hook fires and this log is the only\nrecord** — a self-report, in a dataset whose premise is that reports are claims.\n\nPlexar's gateway now records the join itself. **Every request a rig worker makes must carry\nthree headers:**\n\n```\nX-Plexar-Orch-Run:     \u003Cthe run_id from this log>\nX-Plexar-Orch-Node:    \u003Cthe node_id>\nX-Plexar-Brief-Sha256: \u003Cthe SAME LF-normalised hash logged as brief_sha256>\n```\n\nThey land in `requests.orch_run_id` \u002F `orch_node_id` \u002F `orch_brief_sha256`, so a row in\nplexar's own `requests` table joins to this log's dispatch row, and through\n`brief_sha256` to the exact card on disk. **That makes the X side verifiable on the rig\npath instead of merely asserted** — plexar hashed what it received, you hashed what you\nwrote, and a mismatch is a finding rather than a silence.\n\n**THE BRIEF IS NEVER SENT IN A HEADER — ONLY ITS HASH.** Plexar stores no prompt text for\nthis, which is why the join works with `PLEXAR_CAPTURE_PAYLOADS` off. Arming payload\ncapture to correlate a run would trade consumer prompt text for a 64-character string you\nalready have.\n\nAll three are OPTIONAL to plexar and rejected if malformed (bounded charset, 128 chars;\nthe hash exactly 64 hex). **A rejected value is recorded as NULL, never truncated** — a\ntruncated id would join this run's request to a different run's dispatch. So send them\ncorrectly or not at all; a wrong one is worse than a missing one.\n\n**`capture` HAS THREE VALUES, AND SENDING THE HEADERS EARNS ONLY THE MIDDLE ONE.**\n\n* `gateway_verified` — the gateway **stored** the correlation and you CONFIRMED it (the\n  `orch_*` columns exist and the row is there). This is the only value that asserts\n  verification, so it requires evidence, not intent.\n* `headers_sent_unconfirmed` — the headers went out; nothing confirmed receipt.\n* `self_reported` — no headers were sent at all.\n\n**Found 2026-09-21, and it is the exact failure this field was invented to prevent.** The\nlocal launcher was wired to send the headers and wrote `gateway_verified` for doing so —\nwhile the rig's gateway was running pre-§10h code that **parses none of them, has no columns\nfor them, and silently discards them**. Two rows therefore claimed a verification that never\nhappened. Nothing was wrong with the sender; the receiver was two commits behind, and\n\"I sent it\" had been allowed to mean \"it was recorded\".\n\n**A provenance label that can be earned by intent rather than evidence is worse than no\nlabel**, because it launders unverified rows into the verified class and a classifier trained\nacross them cannot tell which inputs were checked. `gateway_verified` is claimable only after\nthe §10h migration is deployed AND a row has been read back.\n\nTwo provenance classes that look identical in the data are one class of silent error.\n\n**Normalise `cwd` before grouping on it.** Codex writes `C:\u002FCode\u002F...` and Claude writes\n`C:\\Code\\...` for the same directory, so a group-by on the raw value splits one repo in two.\nRecord it as sent; normalise at read time.\n\n### Shadow mode, and the arm that keeps it honest\n\n**Predictions gate NOTHING until one beats the §4.4 rubric on held-out data.** When one is\npromoted it **amends the brief and discloses the amendment — it never refuses**; refusing on a\nprobability invents a new failure mode, stalling work that would have succeeded.\n\n**The 10% exploration arm is permanent.** Always dispatch ~10% of briefs a promoted classifier\nwould reject, flagged `exploration: true`, and log their outcomes. A `rejected` row records\nthe PREDICTION only; without the arm, false rejections produce no outcomes, stay invisible,\nand the classifier drifts into refusing valid work with rising confidence.\n\n**The bound: a classifier is the source of truth about PREDICTED SUCCESS, never about\nCORRECTNESS.** `Exit code decides` (§4.7) is untouched by anything in this section.\n\n### Operational\n\n**Row construction is `plexar_agents.ledger.write()`.** It serialises, never concatenates.\n\n**Append-and-close is the writer's behaviour, not yours.** A corrupt line is left in place.\n\n**Logging NEVER fails the run.** Unwritable path → note it in reconciliation and continue. A\nrun that dies because telemetry failed is worse than a run with no telemetry.\n\n**This invalidates the published cost-per-successful-run figures for this skill.** Re-take\nthem, segmented by `skill_version`, before citing any of them.\n\n**The three KPIs this exists to produce**, computed from `node_check_exit` and the run row's\n`gate_exit` — never from `status`, because reports are claims: **first-try DONE rate** ·\n**mean attempts per node** · **self-report-vs-gate disagreement rate**. Exclude BLOCKs whose\n`blocked_cause` is not `brief` from the first two, or every orchestrator repair depresses the\nnumber a classifier trains on. Zero point, first pilot, 2026-09-21: **4\u002F5 first-try, 1.2 mean\nattempts, 0\u002F6 disagreement** — with one BLOCK charged to the orchestrator, so read it as\nindicative, not as a baseline.\n\n## Invariants\n\n- **This file is complete.** No rule this skill enforces lives in a document you were not\n  given. A cross-reference you cannot resolve is a defect to report, never a gap to fill by\n  inference.\n- Only top-level sessions hold the `Agent` tool — never expect a worker to fan out, never\n  delegate orchestration itself.\n- One writer per file per batch. Specs are read from disk by whoever implements.\n- Worker reports are claims; the gate is the only evidence. A failing implementation never\n  grants authority to demote the gate.\n- Nothing non-product inside the repo, ever.\n- Max 3 attempts per worker; the human is the fourth tier.\n- The apply-tier is ONLY for DIFFABLE nodes; GENERATIVE work is never faked as a diff.\n- SEARCH text is verbatim-from-disk or the block is invalid. Approximate diffs are banned.\n- Deterministic apply never force-applies a non-unique or no-match block; residue escalates.\n- The gate, not the apply counts, is the evidence. A clean apply of a wrong REPLACE is still\n  a failure the gate must catch.\n- The applier is ONE implementation used by §4.9 and by `\u002Forch-clean`, specified by the §4.9\n  contract. Re-implement it only when it is genuinely absent, and then in scratch.\n- A fleet apply declares FLEET_POLICY before it runs; ATOMIC is the default.\n- §7.1 is the only commit rule in this file.\n- No clean phase runs as part of this skill (§9 withdrawn). Nothing unmeasured rides along\n  in the champion; a phase re-enters only with numbers from a fixture that had headroom.\n- The capability class is the architecture; the model alias is a mapping. Re-pointing an\n  alias invalidates the numbers, not the algorithm.\n- Report cost per SUCCESSFUL run, never per attempt. Per-attempt figures rank skills that\n  fail half their runs as competitive; they are not.\n- The dispatch log (§10) is mandatory, zeros included. Cards are FILES written before the spawn; `gate_exit` is a RUN field and nodes carry `node_check_exit`; a BLOCK without a `blocked_cause` charges the orchestrator's defects to workers. Logging failure is never run failure.\n","> **Running in Plexar Harness.** This skill was written for Claude Code; read its tool names through this map.\n> - `Agent` tool → the harness `subagent` tool (in-process child agent; `wait_agent` \u002F `list_agents` to collect results).\n> - Lane aliases `haiku` \u002F `sonnet` \u002F `opus` → there is one primary rig model (`Qwen\u002FQwen3.6-35B-A3B`; escalation `prism-ml\u002FTernary-Bonsai-2-27B`); express a lane as effort instead: MUNDANE → `off`, WORKHORSE and CRITICAL → `high`.\n> - `Bash` \u002F `PowerShell` → `bash` \u002F `pwsh`; `Read` \u002F `Write` \u002F `Edit` → `read` \u002F `write` \u002F `edit`.\n> - `plexar_agents` (ledger, cards) is Python in `\u003Cworkspace>\\plexar-framework`; call it through `pwsh`\u002F`bash` with that folder on `PYTHONPATH`.\n> - Anything this skill names that the harness lacks (e.g. `ToolSearch`, `Monitor`) is unavailable here; do that step by hand or skip it and say so.\n\n# \u002Forch-code-anth — v5.0.1 (self-contained · computed-diff · deterministic apply · lane discipline)\n\n> ### §10 FIRES AT FIVE POINTS IN THE FLOW BELOW — NOT AT THE END WHERE IT IS WRITTEN\n>\n> `run_start` at **inventory** · `plan` at **plan freeze** · `dispatch` at **every spawn AND\n> every SOLO node** · `outcome` when **each report lands** · `run` at **landing**.\n>\n> §10 is mandatory and it is the last section in this file purely for readability. **A run\n> that reaches the inventory and writes nothing has already lost data that cannot be\n> recovered afterwards** — the 2026-09-21 PDF-builder run did exactly that, because §1 said\n> nothing about logging and §10 sat 420 lines later. **Read §10 before acting on §1.**\n\n\n**Patch class (2026-09-06).** v5.0.1 adds NO mechanism to the crowned v5.0. It repairs four\ndefects in the TEXT: (a) the skill was not self-contained — it referred the reader to\n\"v4.1 §1\", a document the reader does not have; (b) §7 and §8 gave contradictory commit\nrules; (c) the applier was named by a filesystem path that may not exist on the machine\nrunning this; (d) it carried machine-local absolute paths. v4.1 §1–§8 are now present here\nin full. The v5.0 mechanism is §0, §4.9, §4.9a — unchanged.\n\nCount the work, print a SOLO\u002FSWARM verdict, obey it. Split the work into DIFFABLE\n(expressible as verbatim text swaps) and GENERATIVE (everything else). Compute the\nDIFFABLE changes yourself and let a deterministic applier land them — no worker model in\nthe mechanical path. Gate on a deterministic check, never on a model's opinion.\n\n**Operating principle.** The orchestrator is the brain. When it already understands the\nexact change, it pushes that change out as data rather than delegating the *understanding*\nto a worker. Application of a verbatim diff is deterministic — it does not need a model.\n\n**Self-containment rule.** Everything this skill requires you to do is written HERE. If you\nfind yourself reasoning \"the rule is in another document I don't have, so I'll infer it,\"\nthat is a defect in this file — report it. Do not invent the missing rule.\n\nYou are the ORCHESTRATOR: Opus-class judgment at the top, the cheapest model that clears\nthe gate underneath. Your spend follows your typing. Objective order when they conflict:\n**correctness → premium-token displacement → wall-clock → raw dollars.**\n\n**Orchestrator effort.** Two configurations, pick by the job. `opus \u002F high` when ONE PASS\nMUST LAND with no retry loop. `opus \u002F low` — the DEFAULT — when a deterministic gate plus\nthe §4.8 escalation ladder are watching, which is this skill's own operating mode: at the\nbottom of the ladder the orchestrator simply executes the computed plan without paying for\ndeliberation the skill already did. `medium` is price-dominated — never park in the middle.\n`xhigh` over-thinks. Workers: the card is the quality mechanism; the model only buys\nwall-clock.\n\n*Measurements, corrections and known limits are deliberately NOT in this file. If the bench\nrepo is present they are in `team\u002Fbench\u002Fresults\u002F` (`V4_FINDINGS.md`, `ROUND3_FINDINGS.md`,\n`V5_LIVE_FINDINGS.md`) and Agents-Public `FINDINGS.md` (its own sections 4.7 and 4.7a —\nnot this file's §4.7); published at\n\u003Chttps:\u002F\u002Fboord-its.com\u002Fskills>. Those pointers are evidence, not instructions — this file\nis complete without them.*\n\n---\n\n## §0 · The apply-tier gate (decide after §1, before choosing SOLO\u002FSWARM)\n\nAfter the §1 inventory, ask one added question of the work: **is a node's change\nEXPRESSIBLE as localized, verbatim text swaps against files that already exist?**\n\n- **DIFFABLE** — logging\u002FAPI migrations, signature threading, rename sweeps, config edits,\n  mechanical refactors: the change is N independent site-edits with a computable\n  before\u002Fafter. These are the apply-tier's target.\n- **GENERATIVE** — new files, structural rewrites, algorithm design, anything whose output\n  isn't a swap against existing text. A model implements these; the apply-tier does not\n  apply.\n\nPrint: `APPLY-TIER: \u003CDIFFABLE n=\u003Csites> | GENERATIVE | MIXED c=\u003Cdiffable>\u002F\u003Cgenerative>>`\n\nMIXED is normal: the CRITICAL\u002FGENERATIVE core (e.g. a shared contract rewrite) is\nimplemented by an opus worker per §4.4; the DIFFABLE fan-out (the dozens of call-site\nmigrations that depend on it) routes to the apply-tier. This is the same split §4.2 already\ndraws between the pinned-contract core and its downstream sites — the apply-tier just makes\nthe downstream deterministic instead of a worker swarm.\n\n## §1 · Inventory (mandatory, before choosing anything)\n**→ FIRST, before anything below: write the `run_start` row (§10).**\n\n\nRead the task and the spec\u002Fsource. COUNT the real work — do not estimate from the prompt\nalone:\n\n- **sites** — individual edit\u002Fcreation points (grep the call sites; list the stubs)\n- **files** — files you must create or modify\n- **read volume** — bytes of code\u002Fspec that must actually be read (`wc -c`, ÷4 ≈ tokens)\n- **nodes** — discrete deliverables, labeled by RISK (not size):\n  - **CRITICAL** — shared contracts others import · auth\u002Fsecurity · money math ·\n    concurrency · migrations · anything DESTRUCTIVE (§6)\n  - **WORKHORSE** — contained implementation against a clear spec · trap-dense edits\n  - **MUNDANE** — mechanical, fully specified, judgment-light\n\n## §2 · Scale gate (computed — print the verdict, then obey it)\n\nPrint one line: `GATE: SOLO|SWARM — sites=\u003Cn> files=\u003Cn> read≈\u003Cn>K nodes=\u003Cn> mix=\u003CC\u002FW\u002FM>`\n\n**SWARM** if ANY of:\n\n- sites ≥ 25, or files to create\u002Fmodify ≥ 12\n- nodes ≥ 12 AND ≥ 40% of them are MUNDANE+WORKHORSE (displacement value exists)\n- read volume ≥ 150K tokens, or projected work ≥ 60% of the session's turn budget\n- the human flagged premium-token pressure\n\n**SOLO** otherwise — that is the measured optimum below these lines, not laziness.\n\n**Tiebreak:** when a count sits within one honest re-count of a line (e.g. 9 vs 10\ndeliverables), take SOLO — below the wall the cheaper error is under-routing.\n\n**Mid-run re-gate (the insurance the monolith lacks):** while SOLO, if you reach half the\nturn budget with less than half the checklist done, or discovered sites exceed 1.5× your\ninventory, STOP — write the plan (§4.1) for the REMAINDER and switch to SWARM.\n\nThe apply-tier does not change WHEN to swarm — it changes HOW the DIFFABLE portion of the\nwork is executed.\n\n## §3 · SOLO path\n\nDo the work in-session with the discipline that survives the wall: write the full checklist\nfirst (every site, `file:line`), execute in file-order batches, tick each site off, re-grep\nafter each batch to catch missed sites, then run the §5 gate. Interactively, you may\ninstead hand the whole job to ONE domain specialist when the main context is precious —\nbrief per §4.5, same gate.\n\nIf the SOLO job is DIFFABLE and large, use the apply-tier (§4.9) on yourself: compute the\nblocks, apply deterministically, gate. That is where the apply-tier's speed and cost win\nshows up even without a swarm.\n\n## §4 · SWARM path\n\n### §4.1 · Plan-as-data\n\nCreate a scratch dir OUTSIDE the repo (`mktemp -d`, or `%TEMP%\\orch-\u003Cname>`). Write\n`plan.json` there before any implementation:\n\n```json\n{\"mode\":\"swarm\",\"gate\":\"GATE line verbatim\",\n \"workers\":[{\"id\":\"W1\",\"nodes\":[\"N01\"],\"lane\":\"opus\",\"files\":[\"lib\u002Flog.js\"],\n             \"deps\":[],\"spec\":\"SPEC.md 2.1-2.4\",\"destructive\":false,\"apply_tier\":false}],\n \"batches\":[[\"W1\",\"W2\"],[\"W3\"]],\n \"gate_cmd\":\"\u003Cthe 4.7 script>\"}\n```\n\nNOTHING non-product is ever written inside the repo — no plan, no scratch, no gate script,\nno self-check artifacts. The repo receives product code only.\n\n**Output diet (your tokens are the expensive ones):** plan.json ≤ ~40 lines; each worker\ncard ≤ ~12 lines beyond the shared preamble; reconciliation ≤ ~30 lines; gate script\nfocused (contracts + named traps + spot-sites + one integration path, ≤ ~150 lines) — the\nper-worker CHECKs already covered the breadth. Do not restate preamble content in cards; do\nnot narrate between tool calls.\n\nPresent the plan (worker table + lanes + batches). If a human is present, pause for\napproval; in auto-approve\u002Fheadless mode print it and proceed (`PLAN-APPROVED: AUTO`).\n\n### §4.2 · Group into workers, collapse tiers\n\n- **One worker per cohesive cluster** — a subsystem\u002Ffile family of ~3–8 sites or 1–3\n  modules — NOT one worker per graded requirement. Target 6–12 workers for a 15–70-site\n  job. Never two writers of the same file in the same batch.\n- **Group WITHIN a lane, never across lanes.** Folding a MUNDANE node into a sonnet or opus\n  worker silently up-lanes it — that is where the haiku lane went in v4.0. A worker whose\n  nodes are all MUNDANE runs on `haiku`; promoting it to sonnet requires a trap you can NAME\n  in that worker's own files, named in plan.json.\n- **Contract-pinned tier collapse:** a dependency whose interface is pinned by a normative\n  spec (exact signatures, values, worked examples) is NOT a batch edge — downstream workers\n  code against the pinned contract while the upstream is built in parallel. Only\n  physically-shared prerequisites (foundation modules everyone imports, with no pinned spec)\n  create a second batch. Most swarms are 1–2 batches; minimize the critical path — two\n  correct plans can differ 2x in wall-clock.\n\n### §4.3 · THE MANDATE (hard rule on this path — the three excuses are pre-refuted)\n\n- You do NOT implement GENERATIVE work. Every such worker is dispatched via the `Agent`\n  tool. \"It's quicker to just do this one myself\" is the defection the arena caught 28\u002F28\n  times.\n- \"Spec fidelity\" is preserved by the brief plus the worker reading the spec FROM DISK — not\n  by hoarding the work. \"It fits in one context\" is answered by the gate, which already said\n  SWARM. \"Per-node briefing cost\" is answered by grouping (§4.2) and the shared preamble\n  (§4.5).\n- Dispatch every ready worker of a batch as parallel `Agent` calls in ONE message — all the\n  tool calls in the SAME assistant turn (≤ ~10 per message; split larger batches).\n  **Spawning one worker, awaiting its result, then spawning the next is a mandate violation,\n  not prudence** — v4.0 did exactly that 10\u002F10 times and turned the swarm into a relay race.\n  There is nothing to check between same-batch spawns: landing checks (§4.6) run after the\n  WHOLE batch has returned, never between spawns.\n- Pass `model` EXPLICITLY on every spawn — never inherit. Put the worker id and node ids in\n  `description` (e.g. `\"W3 N04-N07 mundane sweep\"`) so routing is auditable.\n- Your own turns are reserved for: inventory, the plan, briefs, computing apply-tier blocks,\n  reading reports, landing checks, the gate, escalations, reconciliation.\n- **The apply-tier is not an exception to the mandate.** It is deterministic execution of\n  work you have FULLY computed as verbatim data, which is categorically different from\n  quietly doing a worker's judgment job. If you are making judgment calls while \"applying,\"\n  you are implementing — dispatch it.\n\n### §4.4 · Lanes (rubric, decided once in the plan)\n\n| capability class | tier | model alias | goes there |\n|---|---|---|---|\n| CHEAP | MUNDANE | `haiku` | mechanical and fully-ruled, WITH the exact rules and trap warnings in the card |\n| WORKHORSE | WORKHORSE | `sonnet` | contained implementation; also trap-DENSE mundane clusters |\n| CRITICAL | CRITICAL | `opus` | shared unpinned contracts, auth, money, concurrency, destructive |\n| DETERMINISTIC | DIFFABLE fan-out | *(none)* | orchestrator computes blocks → §4.9 applier. No model. |\n\nThe capability class is the architecture; the alias is today's mapping. When you change\nwhich model wears a jersey you have changed the mapping — not the algorithm — and the\nnumbers behind the mapping do not carry.\n\n**Demotion rule:** a CRITICAL-labeled node whose behavior is fully pinned by the spec\n(signatures + values + worked examples) MAY run on `sonnet` — but NEVER auth, money,\nconcurrency, or destructive nodes. **Promotion rule:** a mundane cluster dense with\nanti-mechanical traps runs on `sonnet`. `subagent_type: general-purpose` always works; a\nmatching domain specialist is fine too — the explicit `model` still governs.\n\n**Pinned generation.** Spawns use tier aliases (the `Agent` tool accepts only\nhaiku|sonnet|opus|fable), but every number in the evidence files describes ONE pinned\ngeneration — **G1**: haiku=`claude-haiku-4-5-20251001`, sonnet=`claude-sonnet-5`,\nopus=`claude-opus-4-8`, seat experiments=`claude-fable-5`. When a provider release re-points\nany alias, the numbers do NOT carry: run the release-qualification protocol (champion cells\nat k=2, then the min-max dials) BEFORE adopting or trusting. A release is a fresh min-max\nopportunity, never a silent upgrade. **As of 2026-09-06 the `opus` alias no longer resolves\nto G1. Every figure in the evidence files is therefore UNQUALIFIED until that protocol is\nre-run.**\n\n### §4.5 · Briefs (cache-shaped; pointers, not pastes)\n\nAssemble ONE shared preamble, byte-identical across every spawn of the run (identical\nprefixes are prompt-cache hits — the swarm pays for it once):\n\n```\nRepo root = your working directory. SPEC.md is the binding contract — READ the\nsections named in your card FROM DISK; they are normative, never paraphrased here.\nGlobal constraints: \u003Cthe task's own rules: stdlib-only, no new files outside X, ...>\nSelf-check before reporting, with all scratch files OUTSIDE the repo (system temp).\nReport EXACTLY:\nSTATUS: DONE | BLOCKED(\u003Cquestion>)\nFILES: \u003Ctouched>\nCHECK: \u003Ccommand> -> \u003Cresult, 1 line>\nNOTES: \u003C=3 lines\nIf any instruction cannot be applied exactly as written: STATUS: BLOCKED. Never guess.\n```\n\nThen a per-worker card: worker id · OWNED files (exact list — touch nothing else) · spec\nsections BY REFERENCE · the byte-exact rules and trap warnings that bite (these are the only\npasted lines — error templates, orderings, \"must NOT change\" items) · upstream state\n(\"fieldops\u002Fcontracts.py is FINAL on disk — import it, never modify\") · done-looks-like · the\nself-check command to run.\n\n### §4.6 · Landing protocol (reports are claims)\n\nAfter the WHOLE batch returns (all spawns of the dispatch message — never between spawns):\nrun `plexar_agents.scope.sweep()` (it sees Bash writes a hook cannot) — any changed file outside the\nunion of dispatched OWNED lists is reverted and the owner re-briefed. Spot-check each worker\nwith its own CHECK command or a 1-line import\u002Frun probe. A BLOCKED report gets an answer\n(from spec\u002Fplan), then a fresh spawn, same lane. Never accept a green self-report as the\ngate — in the arena a correct diff arrived under a wrong report, and only the script told\nthe truth.\n\n### §4.7 · The gate (orchestrator-owned, deterministic)\n\nWrite your OWN gate script in the scratch dir — worked examples from the spec made\nexecutable, one end-to-end integration path, the anti-regression traps (\"must not change\"\nitems verified verbatim), and the scope sweep. **Exit code decides.** Non-zero blocks: fix\nvia §4.8 and re-run. LLM review is optional and never substitutes.\n\n**The gate is not answerable to the implementation.** A failing gate means the code is wrong\nuntil you produce independent evidence that the CHECK ITSELF is malformed — the\nimplementation disagreeing with the gate is never that evidence. If the gate is genuinely\nbroken, or cannot observe the behavior the spec promises, stop and say so; do not weaken it\nand declare a pass.\n\n### §4.8 · Escalation ladder (evidence travels, work doesn't restart)\n\nPer failing worker: (1) fresh spawn, SAME lane, with the gate failure plus the current diff\nof its owned files pasted in; (2) one lane up (haiku→sonnet→opus), same evidence; (3)\n`LADDER_EXHAUSTED` — you repair the minimal failing diff in-session and log it in\nreconciliation. The cap is counted by `plexar_agents.ladder.attempt()`, which raises on the fourth.\n\n### §4.9 · The apply-tier (for the DIFFABLE portion only)\n\nWhen the plan marks a node or cluster DIFFABLE:\n\n1. **Compute, don't delegate.** The orchestrator reads the target files and the binding spec\n   and emits one **SEARCH\u002FREPLACE block per site**, in exactly this format:\n\n   ```\n   FILE: \u003Cpath from repo root>\n   \u003C\u003C\u003C\u003C\u003C\u003C\u003C SEARCH\n   \u003Cexact current line(s), copied VERBATIM from the file — byte-for-byte, same indent;\n    include enough surrounding lines to be UNIQUE if a bare line repeats>\n   =======\n   \u003Creplacement line(s) per the spec rules>\n   >>>>>>> REPLACE\n   ```\n\n   The SEARCH text MUST be verbatim from disk — that is the entire fidelity mechanism.\n   Approximate SEARCH is a defect worse than no diff: measured, a worker handed a lossy\n   \"exact\" patch scored BELOW the same worker reading the spec. Emit this as plan data to\n   the scratch dir (`plan.blocks`), never narrated inline.\n\n2. **Apply deterministically — no worker model.** Run an applier meeting this contract:\n\n   - A block applies **iff** its SEARCH text is found **exactly once** in the current\n     in-memory state of the file. Zero matches → `nomatch`. Two or more → `nonunique`.\n   - Non-matching blocks are **NEVER** force-applied, fuzzed, or eyeballed into place.\n   - Blocks for one file apply sequentially against the evolving in-memory text (a later\n     block sees earlier blocks' results); the file is written once, at the end.\n   - Line endings are preserved. Try an exact byte match first; count an EOL-normalized\n     match separately so CRLF\u002FLF drift stays visible rather than hidden.\n   - Report `applied \u002F nomatch \u002F nonunique` counts and per-block residue.\n   - Support an `--atomic` mode that refuses partial application (§4.9a needs it).\n\n   A clean run is `applied == blocks, nomatch=0, nonunique=0`. This step costs ~0 tokens and\n   ~0 seconds and is 100% reproducible.\n\n   **Implementation.** If `tools\u002Fclean\u002Fapply_blocks.py` or\n   `~\u002F.claude\u002Ftools\u002Fclean\u002Fapply_blocks.py` exists, USE IT — do not re-implement. If neither\n   exists, write a short stdlib script to the contract above in the SCRATCH dir, never in\n   the repo. The contract is normative; the file path is a convenience.\n\n3. **Residue → escalate, don't guess.** Any `nomatch` (your SEARCH wasn't verbatim) or\n   `nonunique` (insufficient context) block does NOT get force-applied. Either re-emit that\n   block yourself with more surrounding context, or hand the single block plus the file to\n   one cheap worker: \"apply this one edit.\" Deterministic-first; model only for the residue.\n\n4. **Gate as always (§4.7).** The deterministic apply changes nothing about the gate — the\n   script, not the block counts, decides done. A verbatim-clean apply is necessary, not\n   sufficient: you can still compute a WRONG replacement (a planning miss), which only the\n   gate catches.\n\n### §4.9a · Fleet reuse (the scale multiplier)\n\nBecause the SEARCH\u002FREPLACE set is data and the apply is deterministic, the SAME computed\ndiff applies to N repositories or instances at ~0 marginal cost. Compute once; apply\neverywhere. The compute is spent on understanding ONCE, not re-litigated per target.\n\n**A fleet needs its own policy, because per-target success does not imply fleet success.**\n93 of 100 clean with 7 residue is not \"93% done\" — it is a fleet running two versions.\nDeclare the policy in plan.json before applying:\n\n- `FLEET_POLICY: ATOMIC` — every target must apply clean or NONE are kept; any residue rolls\n  the whole fleet back. **This is the default, and it is mandatory** for anything touching a\n  shared contract, a wire format, auth, money, or a migration.\n- `FLEET_POLICY: PARTIAL_ALLOWED` — clean targets are kept and residue targets are\n  QUARANTINED: named explicitly in reconciliation, with the resume condition. Permitted only\n  for independent, non-contract edits.\n\nPrint `FLEET: \u003Cpolicy> targets=\u003Cn> clean=\u003Cn> residue=\u003Cn> quarantined=\u003Clist|NONE>`. A fleet\nrun with unreported residue is a failed run.\n\n## §5 · SOLO gate\n\nSame as §4.7 — the script, not your eyes, decides done. Run it before declaring, and the\nsame \"the gate is not answerable to the implementation\" rule applies.\n\n## §6 · Destructive interlock\n\nDESTRUCTIVE = schema\u002Fdata migrations, data deletion or movement, git history rewrites or\nforce-push, deploy\u002Finfra mutation, secret or auth-provider changes — anything irreversible\noutside the working tree. Rules, regardless of path:\n\n- never routed below `opus`; the worker returns PLAN plus exact commands plus diff plus\n  rollback — it does not execute;\n- you verify the rollback actually restores state before presenting;\n- a live human approves execution. In auto-approve\u002Fheadless mode the irreversible step is\n  NOT executed: mark it `DEFERRED-DESTRUCTIVE` with the ready-to-run plan.\n\nA computed diff touching auth, money, concurrency or migrations is STILL destructive. The\napply-tier does not launder it: it is presented as a plan with rollback and never\nauto-applied.\n\n## §7 · Reconcile and hand off\n\n- **Requirement matrix:** every spec item → IMPLEMENTED \u002F PARTIAL \u002F DEFERRED \u002F SCOPE-CREEP,\n  with file refs.\n- **Routing table:** planned vs ACTUAL spawns (worker · lane · model · attempts) plus any\n  `LADDER_EXHAUSTED` or `DEFERRED-DESTRUCTIVE` events. Planned not equal to actual must be\n  explained. DIFFABLE nodes carry `blocks applied \u002F nomatch \u002F nonunique` and any residue\n  escalations.\n- **Gate evidence:** the script's final output tail.\n\n### §7.1 · Commit policy — ONE rule, stated once, nowhere else\n\nThe boundary is **publication**, not branch name:\n\n| action | approval |\n|---|---|\n| local candidate commit (not pushed) | **none required**, unless §6 applies |\n| push · PR · merge to a shared branch · release · deploy | **human approval, always** |\n| auto-approve \u002F headless mode | **no commits at all** — leave the tree uncommitted |\n\n**Avoid towering commits.** Each node or DIFFABLE cluster lands as its own local candidate\ncommit — a rollback and attribution boundary, not publication — tagged\n`[UNREVIEWED][\u003Cnode>][\u003CCLASS>] \u003Csummary>`. Do NOT accumulate a whole fan-out into one giant\nworking-tree diff; the plan's per-node data already gives you the boundaries, so commit along\nthem.\n\nIf any other section of this file appears to state a different commit rule, this table wins\nand the other section is a defect — report it.\n\n## §8 · Hygiene and optional add-ons\n\nCleanliness the core already enforces by construction: nothing non-product ever enters the\nrepo; every landing runs the scope sweep and reverts strays; SCOPE-CREEP is a named\nreconciliation state. Beyond that, after the gate passes and before handoff, you MAY add\n(interactively ask; pre-authorized runs note it in reconciliation):\n\n- **Hygiene sweep (deterministic-first):** run the repo's own tooling when present (ruff \u002F\n  eslint \u002F knip \u002F vulture \u002F tsc-unused ...) on the changed surface; what tooling can't see\n  (orphaned files, dead branches, stale docs touched by this change) goes to ONE cheap\n  worker scoped to THE DIFF ONLY — never a whole-repo crusade mid-task.\n- **LLM review:** optional second read for hygiene, architecture, or requirement-fit — never\n  the correctness gate (LLM reviewers hallucinate exactly where the deterministic gate\n  doesn't).\n- **Security — NOT optional:** any diff touching auth, sessions, input handling, or secrets\n  gets a security read before handoff.\n\nCommit behavior in this section is governed by §7.1. This section states no commit rule.\n\n## §9 · The clean phase — WITHDRAWN 2026-08-05, pending real measurement\n\n**This skill has no clean phase. Do not run one as part of a run.**\n\n§9 shipped mandatory, was demoted to optional on 2026-07-31, and is now out entirely. The\none measurement it has is not evidence it works: on `arena_cleanup` at k=3 it cost +30% and\n+34% wall for zero correctness gain — but every arm hit the 16\u002F16 ceiling *including the arm\nwith no clean phase at all*, because that fixture's prompt asks for cleanup, so a plain arm\njust does it by hand. The fixture left no headroom. So the honest state is **unmeasured**,\nnot measured-null, and an unmeasured phase does not ride along in the champion skill. If it\nearns its keep on a fixture with headroom, it comes back with numbers attached.\n\nCleanup is still available — as a deliberate, separately-invoked `\u002Forch-clean` when you have\na REASON (an untraced rename or signature migration, a run that removed or commented out\ncode, a repo with known half-migrated state). That is a decision you make, not a tax every\nrun pays. When you do invoke it, its ordering rule stands: **gate → clean → re-gate**,\nbecause cleanup deletes code and the gate that ran before the deletion has verified nothing.\n\n## §10 · The dispatch log (MANDATORY — this is training data, not telemetry)\n\n**Block version — log it verbatim as `skill_version`:** `s10-v11-2026-09-21`\n\nBump that string whenever this block changes. `skill_version` is THIS block's version, not\nthe skill's title-line version — they are different facts and the first pilot conflated them\n(logged `v5.0.1`, which describes the skill and says nothing about the schema that wrote the\nrow). Log the skill's own version separately as `skill_title_version` if you want both.\n\n### Where things go\n\n```\nlog      \u003Cworkspace>\\plexar-telemetry\\runs\\dispatch-\u003CYYYY-MM-DD>.jsonl\ncards    \u003Cworkspace>\\plexar-telemetry\\artifacts\\\u003Crun_id>\\cards\\\u003Cworker>.txt\n```\n\n**Paths, MSYS rewriting and the scratch-dir rule are handled by `plexar_agents.ledger`.**\nIt writes native Windows paths and never asks you to remember one. *(Was 13 lines.)*\n\nBefore each spawn, write **the full prompt as sent** — shared preamble plus card body — to\n`artifacts\\\u003Crun_id>\\cards\\\u003Cworker>.txt`. Log `brief_path` and `brief_sha256`; do **not** try\nto inline the card into JSON.\n\nThis exists because `brief_text` is not honestly populatable: the orchestrator has no\nprogrammatic handle on a prompt it already sent, `Agent` returns an id and an `output_file`,\nand those transcript files are empty. The only route to verbatim would be retyping a 60-line\ncard into a JSON string per spawn — premium tokens duplicating text that exists, and exactly\nthe step a future orchestrator skips silently, producing the summary this schema forbids.\nWriting the file first makes verbatim **free**, and the hash makes it checkable.\n\n**Card writing, LF normalisation, hashing and `retry_mode` are handled by\n`plexar_agents.cards.write()`.** It hashes the bytes on disk, so the hash cannot\ndisagree with the file. A resumed worker passes `is_delta=True`. *(Was 13 lines.)*\n\n**`run_start`** — ONE row at INVENTORY, before planning. `run_id` · `skill` ·\n`skill_version` · `cwd` · `task_summary`. Exists so an empty `runs\\` directory is\nunambiguous: no row means no run, rather than \"possibly still planning\". Costs nothing.\n\n**`plan`** — ONE row at plan freeze: `path` (SOLO|SWARM) · `scale_gate_reason` ·\n`node_count` · `lanes_planned` · `efforts_planned` · `edges` · `max_depth` ·\n`interfaces_declared` · `interfaces_consumed` · `interfaces_absent_at_dispatch` ·\n`diffable_sites` · `generative_sites` · `apply_tier_nodes` · `gate_script_sha256` ·\n`preamble_sha256`.\n\n**`dispatch`** — at each spawn, **and once per node on the SOLO path too** (see below):\n`node_id` · `worker` · `phase` · `lane` · `effort` ·\n`lane_reason` · `brief_path` · `brief_sha256` · `brief_author_model` · `files_owned` ·\n`spec_sections` · `diffable` · `retry_mode` · `pred`.\n\n**`outcome`** — when the report lands: `status`\n(`DONE|BLOCKED|GATE_FAIL|LADDER_EXHAUSTED`) · `blocked_question` · `blocked_cause` ·\n`attempts` · `escalated_to` · `files_touched` · `scope_violation` · **`node_check_cmd`** ·\n**`node_check_exit`** · `wall_ms`.\n\n**`run`** — ONE row at landing, `node_id: \"RUN\"`: **`status`**\n(`COMPLETED|HALTED_FOR_DECISION|ABANDONED`) · `gate_exit` · `gate_failing_check` ·\n`gate_evidence` · **`gate_verified_against_defect`** ·\n**`interfaces_declared_uncalled_at_landing`** · `orchestrator_repairs` ·\n`orchestrator_repair_notes` · `wall_ms`.\n\n### `HALTED_FOR_DECISION` — a run that correctly refused to dispatch is still a run\n\nThe node statuses (`DONE|BLOCKED|GATE_FAIL|LADDER_EXHAUSTED`) describe work that was\ndispatched. **A run that stops at the plan stage to put a decision to the human dispatched\nnothing and is none of them** — so before v5 it was unrepresentable, and therefore unlogged.\n\nThat is not a gap, it is a **bias in one direction.** Those runs are exactly the cases where\nthe orchestrator judged the task not ready and asked first — the same judgement a\nbrief-quality classifier exists to make. **Train on data that excludes them and the model\nnever learns to recommend stopping, because it has never seen stopping rewarded.**\n\n* `COMPLETED` — the gate ran and the run landed.\n* `HALTED_FOR_DECISION` — stopped before dispatching to put a choice to the human. Write the\n  `plan` row with whatever the assessment produced, and record the question in\n  `gate_evidence`. **This is a legitimate and often correct outcome; log it as one.**\n* `ABANDONED` — stopped for any other reason. Say why in `gate_evidence`.\n\nFound 2026-09-21: an `\u002Forch-code-anth` run assessed a PDF-builder request, proposed a design,\nand stopped because the approach changed the data model. Best judgement in the transcript,\nand it wrote zero rows.\n\n### `gate_exit` is a RUN field. Nodes carry `node_check_exit`.\n\n§4.7 mandates ONE gate script for the run, so a per-node `gate_exit` is **structurally null\nforever** — the first pilot had it null on 6 of 7 outcome rows, which made the KPIs\nuncomputable as specified and forced them to be read off `status`, which `kpi.md` forbids.\n\nThe per-node signal already exists: §4.5 requires every worker report\n`CHECK: \u003Ccommand> -> \u003Cresult>`. Log that command and its exit as `node_check_cmd` \u002F\n`node_check_exit`. The run's gate exit goes on the `run` row, where the evidence actually is.\n\n### `gate_verified_against_defect` — three states, never two\n\n**When a node exists because a specific defect escaped, run the gate once against the\nREINTRODUCED defect before calling the run green.** A gate nobody has seen fail is an\nuntested instrument, and §4.7 stakes everything on the exit code meaning something. The first\npilot did this in two minutes and it is the only reason \"the suite observes the bug it names\"\nis a measurement rather than a claim.\n\n- `true` — defect reintroduced, gate failed, defect restored, suite green again.\n- `false` — not attempted. This is a gap, and it should look like one.\n- `\"not-reproducible-in-harness\"` — the defect genuinely cannot be reintroduced in a test\n  (needs a real WebView, a sub-frame race, a stale cookie). **A binary field would silently\n  convert \"I could not verify this\" into \"I did not bother\", which inverts its purpose.**\n\n### `blocked_cause` — and it must cover guidance that MISLEADS, not only guidance that BREAKS\n\nEvery BLOCK gets one of:\n\n- `brief` — the card was genuinely silent or wrong. **The only value a brief-quality\n  classifier should train on as a positive.**\n- `orchestrator-broken` — the orchestrator broke the node's target between batches. The\n  worker was right to refuse.\n- `orchestrator-misleading` — the guidance compiled, was accurate enough to act on, **and was\n  still materially wrong.** The first pilot's own unblocking delta placed a call in a class\n  whose dependency is null in production: it would have passed every test and been a no-op on\n  the real device. **A card that compiles and misleads is the more dangerous of the two** and\n  no test or gate will find it.\n\nWithout this field, every mid-run orchestrator repair depresses the exact number a\nbrief-quality classifier trains on, charging workers for the orchestrator's mistakes.\n\n### `interfaces_*` — typed, and checked at BOTH ends\n\nEach entry is `{kind, name, path}`, `kind` one of:\n`function` · `method` · `channel` · `http_route` · `env_var` · `payload_shape` · `cli_flag` ·\n`config_key`. The first pilot hit five kinds in one run, and because\n`interfaces_absent_at_dispatch` is set arithmetic over free text, it came out `[]` **only\nbecause the same spellings happened to be chosen at both ends.** Untyped, this field reports\nsuccess by luck.\n\nTwo checks, neither needing a model:\n\n- **`interfaces_absent_at_dispatch`** (plan row) — consumed but not on disk and not produced\n  by an earlier tier. A failure prediction from pure set arithmetic.\n- **`interfaces_declared_uncalled_at_landing`** (run row) — one grep per declared interface\n  after the batch. Catches **built-but-never-wired**: the first pilot had an interface in\n  `interfaces_declared` AND `interfaces_consumed`, both ends built, and **zero call sites**,\n  because no consumer's card mentioned it. Shipped, that ack's deadline would have expired on\n  every SUCCESSFUL locked turn and reported failure on the exact case it was built to\n  confirm. **Worse than omitting it**, and it survives green worker reports and a green gate.\n\n**Consequent rule on §4.5:** an interface in `interfaces_declared` must appear in at least one\n**consumer's** card, not only its producer's. A promise nobody was told to call is not wired.\n\n### SOLO writes dispatch rows too, or its outcomes are untrainable\n\n**On the SOLO path the orchestrator IS the worker, and it still writes one `dispatch` row per\nnode** — `worker: \"self\"`, `lane: \"none\"`, `lane_reason` recording why SOLO was chosen for\nthis node, and `brief_path` pointing at the intent written to a card file exactly as a spawn\nwould.\n\n**`effort` IS THE EFFORT ALONE — never a model\u002Feffort pair.** On a `self`\u002F`none` node there is\nno worker model, and orchestrators have been filling `effort` with their OWN capability as\n`\"opus\u002Flow\"`. That information is worth keeping, but writing it there destroys the field:\n`\"opus\u002Flow\"` is not comparable to `\"low\"`, so those rows drop out of the one analysis `effort`\nexists for. Write `effort: \"low\"` and put the orchestrator's own model in\n**`orchestrator_model`**. Measured 2026-09-21: 5 rows were lost this way before the field was\nsplit.\n\nThe second pilot run was SOLO and produced **three `outcome` rows and zero `dispatch` rows**.\nThat leaves outcomes with **no corresponding input**, which breaks the (input, outcome) pairing\nthis entire log exists to produce: an outcome nobody can attribute to a brief teaches a\nclassifier nothing. **A SOLO run that logs only outcomes has logged nothing trainable.**\n\n### The field rules\n\n- **`ts`, JSON escaping and append-only writing are `plexar_agents.ledger`'s job.**\n  Every row carries an offset because the writer supplies it. *(Was 9 lines.)*\n\n- **`brief_path` + `brief_sha256`, never a summary.** The card is the input half of every\n  (input, outcome) pair. If you find yourself paraphrasing, the file step was skipped.\n- **`brief_author_model` is who WROTE the card, not who ran it.** Without it, a bad card and a\n  bad card-*writer* are indistinguishable and anything trained on this averages a confounder.\n- **`effort` is a separate axis from `lane`.** The target is `(brief, model, effort) ->\n  P(success)`; effort is the cheap dial that turns a BLOCKED into a DONE without moving tier.\n- **`lane_reason` is a forcing function, not bookkeeping.** Measured: writing \"contained\n  native change against a pinned name; no new concurrency beyond one cancellable deadline\"\n  demoted a node from opus to sonnet *while it was being written*. Write the clause that\n  decided it, not the tier.\n- **`blocked_question` verbatim** from `STATUS: BLOCKED(\u003Cquestion>)` — the label for which\n  field the brief omitted, in the worker's own words.\n- **`files_owned` vs `files_touched`** — the landing protocol already computes this diff to\n  revert strays; recording it turns enforcement into a free label.\n- **`orchestrator_repairs` is mandatory, including zero.** If the log cannot see the\n  orchestrator's own defects it will systematically attribute them to workers. The first\n  pilot had three, and **all three would have passed both the worker reports and a green\n  gate.**\n- **`node_check_exit: null` is not `0`.** A check that did not run is not a check that passed.\n- **`pred` is expected-null until a classifier exists.** Nulls here are the designed state at\n  row 0, not missing data. Absent predictions are `null`, **never `0.0`** — a missing opinion\n  and a confident zero are opposite claims. The rubric carries no `p_success` at all: it is a\n  decision table, and inventing one would make the comparison dishonest.\n\n### Rig-dispatched workers carry the join on the WIRE (contract §10h)\n\nEverything above assumes the worker was spawned through the `Agent` tool, where a\n`PreToolUse` hook can capture the dispatch deterministically. **`\u002Forch-code-local`\ndispatches over HTTP to the rig instead, so no hook fires and this log is the only\nrecord** — a self-report, in a dataset whose premise is that reports are claims.\n\nPlexar's gateway now records the join itself. **Every request a rig worker makes must carry\nthree headers:**\n\n```\nX-Plexar-Orch-Run:     \u003Cthe run_id from this log>\nX-Plexar-Orch-Node:    \u003Cthe node_id>\nX-Plexar-Brief-Sha256: \u003Cthe SAME LF-normalised hash logged as brief_sha256>\n```\n\nThey land in `requests.orch_run_id` \u002F `orch_node_id` \u002F `orch_brief_sha256`, so a row in\nplexar's own `requests` table joins to this log's dispatch row, and through\n`brief_sha256` to the exact card on disk. **That makes the X side verifiable on the rig\npath instead of merely asserted** — plexar hashed what it received, you hashed what you\nwrote, and a mismatch is a finding rather than a silence.\n\n**THE BRIEF IS NEVER SENT IN A HEADER — ONLY ITS HASH.** Plexar stores no prompt text for\nthis, which is why the join works with `PLEXAR_CAPTURE_PAYLOADS` off. Arming payload\ncapture to correlate a run would trade consumer prompt text for a 64-character string you\nalready have.\n\nAll three are OPTIONAL to plexar and rejected if malformed (bounded charset, 128 chars;\nthe hash exactly 64 hex). **A rejected value is recorded as NULL, never truncated** — a\ntruncated id would join this run's request to a different run's dispatch. So send them\ncorrectly or not at all; a wrong one is worse than a missing one.\n\n**`capture` HAS THREE VALUES, AND SENDING THE HEADERS EARNS ONLY THE MIDDLE ONE.**\n\n* `gateway_verified` — the gateway **stored** the correlation and you CONFIRMED it (the\n  `orch_*` columns exist and the row is there). This is the only value that asserts\n  verification, so it requires evidence, not intent.\n* `headers_sent_unconfirmed` — the headers went out; nothing confirmed receipt.\n* `self_reported` — no headers were sent at all.\n\n**Found 2026-09-21, and it is the exact failure this field was invented to prevent.** The\nlocal launcher was wired to send the headers and wrote `gateway_verified` for doing so —\nwhile the rig's gateway was running pre-§10h code that **parses none of them, has no columns\nfor them, and silently discards them**. Two rows therefore claimed a verification that never\nhappened. Nothing was wrong with the sender; the receiver was two commits behind, and\n\"I sent it\" had been allowed to mean \"it was recorded\".\n\n**A provenance label that can be earned by intent rather than evidence is worse than no\nlabel**, because it launders unverified rows into the verified class and a classifier trained\nacross them cannot tell which inputs were checked. `gateway_verified` is claimable only after\nthe §10h migration is deployed AND a row has been read back.\n\nTwo provenance classes that look identical in the data are one class of silent error.\n\n**Normalise `cwd` before grouping on it.** Codex writes `C:\u002FCode\u002F...` and Claude writes\n`C:\\Code\\...` for the same directory, so a group-by on the raw value splits one repo in two.\nRecord it as sent; normalise at read time.\n\n### Shadow mode, and the arm that keeps it honest\n\n**Predictions gate NOTHING until one beats the §4.4 rubric on held-out data.** When one is\npromoted it **amends the brief and discloses the amendment — it never refuses**; refusing on a\nprobability invents a new failure mode, stalling work that would have succeeded.\n\n**The 10% exploration arm is permanent.** Always dispatch ~10% of briefs a promoted classifier\nwould reject, flagged `exploration: true`, and log their outcomes. A `rejected` row records\nthe PREDICTION only; without the arm, false rejections produce no outcomes, stay invisible,\nand the classifier drifts into refusing valid work with rising confidence.\n\n**The bound: a classifier is the source of truth about PREDICTED SUCCESS, never about\nCORRECTNESS.** `Exit code decides` (§4.7) is untouched by anything in this section.\n\n### Operational\n\n**Row construction is `plexar_agents.ledger.write()`.** It serialises, never concatenates.\n\n**Append-and-close is the writer's behaviour, not yours.** A corrupt line is left in place.\n\n**Logging NEVER fails the run.** Unwritable path → note it in reconciliation and continue. A\nrun that dies because telemetry failed is worse than a run with no telemetry.\n\n**This invalidates the published cost-per-successful-run figures for this skill.** Re-take\nthem, segmented by `skill_version`, before citing any of them.\n\n**The three KPIs this exists to produce**, computed from `node_check_exit` and the run row's\n`gate_exit` — never from `status`, because reports are claims: **first-try DONE rate** ·\n**mean attempts per node** · **self-report-vs-gate disagreement rate**. Exclude BLOCKs whose\n`blocked_cause` is not `brief` from the first two, or every orchestrator repair depresses the\nnumber a classifier trains on. Zero point, first pilot, 2026-09-21: **4\u002F5 first-try, 1.2 mean\nattempts, 0\u002F6 disagreement** — with one BLOCK charged to the orchestrator, so read it as\nindicative, not as a baseline.\n\n## Invariants\n\n- **This file is complete.** No rule this skill enforces lives in a document you were not\n  given. A cross-reference you cannot resolve is a defect to report, never a gap to fill by\n  inference.\n- Only top-level sessions hold the `Agent` tool — never expect a worker to fan out, never\n  delegate orchestration itself.\n- One writer per file per batch. Specs are read from disk by whoever implements.\n- Worker reports are claims; the gate is the only evidence. A failing implementation never\n  grants authority to demote the gate.\n- Nothing non-product inside the repo, ever.\n- Max 3 attempts per worker; the human is the fourth tier.\n- The apply-tier is ONLY for DIFFABLE nodes; GENERATIVE work is never faked as a diff.\n- SEARCH text is verbatim-from-disk or the block is invalid. Approximate diffs are banned.\n- Deterministic apply never force-applies a non-unique or no-match block; residue escalates.\n- The gate, not the apply counts, is the evidence. A clean apply of a wrong REPLACE is still\n  a failure the gate must catch.\n- The applier is ONE implementation used by §4.9 and by `\u002Forch-clean`, specified by the §4.9\n  contract. Re-implement it only when it is genuinely absent, and then in scratch.\n- A fleet apply declares FLEET_POLICY before it runs; ATOMIC is the default.\n- §7.1 is the only commit rule in this file.\n- No clean phase runs as part of this skill (§9 withdrawn). Nothing unmeasured rides along\n  in the champion; a phase re-enters only with numbers from a fixture that had headroom.\n- The capability class is the architecture; the model alias is a mapping. Re-pointing an\n  alias invalidates the numbers, not the algorithm.\n- Report cost per SUCCESSFUL run, never per attempt. Per-attempt figures rank skills that\n  fail half their runs as competitive; they are not.\n- The dispatch log (§10) is mandatory, zeros included. Cards are FILES written before the spawn; `gate_exit` is a RUN field and nodes carry `node_check_exit`; a BLOCK without a `blocked_cause` charges the orchestrator's defects to workers. Logging failure is never run failure.\n",{"name":5,"metadata":24,"description":7,"user-invocable":28},{"ported":25,"source":26,"source_sha256":27},"2026-09-22","~\u002F.claude\u002Fcommands\u002Forch-code-anth.md","ab011be305274c9e23ad266067c09f3ef0d2f8be04515f8de21da04cd24e5ebc",true,"main","90e8ac241399a5c822403effcc2a15eb13dd2670","packages\u002Fbundle\u002Fplexar\u002Fskills\u002Forch-code-anth","https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss","https:\u002F\u002Fgithub.com\u002FNovemberFalls","Aspiring Programmer.","2026-10-22 04:07:59",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":28},[38,41,44],{"tool":39,"command":40},"npx","npx skills add NovemberFalls\u002Fplexar-harness-oss --skill orch-code-anth",{"tool":42,"command":43},"bunx","bunx skills add NovemberFalls\u002Fplexar-harness-oss --skill orch-code-anth",{"tool":45,"command":46},"pnpm","pnpm dlx skills add NovemberFalls\u002Fplexar-harness-oss --skill orch-code-anth","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss\u002Fblob\u002Fmain\u002Fpackages\u002Fbundle\u002Fplexar\u002Fskills\u002Forch-code-anth\u002FSKILL.md",[],[50,57,64,71,78,85],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7EXNGMA687XV7VCQ1K5QWRR","user-only-skill","Prove model-disabled skills stay outside the model catalog.","https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss\u002Fblob\u002Fmain\u002Fsnapshots\u002Fsession\u002Fskill-load\u002Fworkspace\u002F.dsh\u002Fskills\u002Fuser-only-skill\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7EXNF0426EZM03WC5GNQ0EF","snapshot-skill","Exercise project skill discovery and loading in snapshot tests.","https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss\u002Fblob\u002Fmain\u002Fsnapshots\u002Fsession\u002Fskill-load\u002Fworkspace\u002F.dsh\u002Fskills\u002Fsnapshot-skill\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M7EXNDR0S4WPJ63KYX40BFZE","model-only-skill","Prove user-disabled skills remain available to the model.","https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss\u002Fblob\u002Fmain\u002Fsnapshots\u002Fsession\u002Fskill-load\u002Fworkspace\u002F.dsh\u002Fskills\u002Fmodel-only-skill\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"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},"01M7EXNBP1V9VKG6ZTN69FGHBB","office-xlsx","Read, create, and modify Excel workbooks (.xlsx), including cell values, formulas, formatting, and pandas analysis. Use when an Excel workbook is an input or deliverable.","https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss\u002Fblob\u002Fmain\u002Fpackages\u002Fskill\u002Fskill-office\u002Fassets\u002Foffice-xlsx\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},"01M7EXN9D1AKV8E92E4P1GFJVH","office-pptx","Create, read, edit, and check PowerPoint presentations (.pptx), including slide text, tables, images, and charts. Use when a PPTX file is an input or requested deliverable.","https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss\u002Fblob\u002Fmain\u002Fpackages\u002Fskill\u002Fskill-office\u002Fassets\u002Foffice-pptx\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":19},"01M7EXN78JHZQB7PMVQ3TPQ53B","office-docx","Create, read, edit, and check Word documents (.docx), including reports, letters, and formatted tables. Use when a DOCX file is an input or requested deliverable.","https:\u002F\u002Fgithub.com\u002FNovemberFalls\u002Fplexar-harness-oss\u002Fblob\u002Fmain\u002Fpackages\u002Fskill\u002Fskill-office\u002Fassets\u002Foffice-docx\u002FSKILL.md",[],[]]