[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78GR26XMDX0KAF5MYZ2W0TN":3,"skill-same:01M78GR26XMDX0KAF5MYZ2W0TN":57},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":14,"repositoryHtmlUrl":27,"repositoryHomepageUrl":28,"ownerHtmlUrl":29,"ownerBlogUrl":30,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M78GR26XMDX0KAF5MYZ2W0TN","spec",null,"Trigger immediately when the user says any of: 'spec it out', 'spec this out', 'before you build', 'before we build', 'before implementation', 'let's discuss before building', 'discuss and spec', 'PRD this', 'shape this', 'product spec', 'lightweight spec'. Also use for brainstorming, vague tickets, product shaping, or writing a lightweight product spec\u002FPRD when the problem, scope, desired behavior, or success criteria are unclear. Do not skip this skill and answer conversationally yourself — if a trigger phrase appears, route here even if the request seems clear or auto mode is active. Do not use for implementation design when requirements are already clear; use blueprint instead.","sandsower\u002Fnopal","sandsower","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F302322?v=4","01M78GNBK4G12AGP2PA2FND7R4","Vic Valenzuela","01M78GNBM80RBPX57WYVHFHC97",0,"MIT","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fspec\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: spec\ndescription: \"Trigger immediately when the user says any of: 'spec it out', 'spec this out', 'before you build', 'before we build', 'before implementation', 'let's discuss before building', 'discuss and spec', 'PRD this', 'shape this', 'product spec', 'lightweight spec'. Also use for brainstorming, vague tickets, product shaping, or writing a lightweight product spec\u002FPRD when the problem, scope, desired behavior, or success criteria are unclear. Do not skip this skill and answer conversationally yourself — if a trigger phrase appears, route here even if the request seems clear or auto mode is active. Do not use for implementation design when requirements are already clear; use blueprint instead.\"\n---\n\n# Spec\n\nShape an idea, vague ticket, or open product question into a lightweight product spec. This is the product\u002Frequirements gate, not the implementation-design gate.\n\nUse this for:\n- Open-ended brainstorming\n- Vague tickets that are not ready for implementation design\n- Product behavior, scope, and acceptance criteria discovery\n- Writing a lightweight spec\u002FPRD before `blueprint`\n\nDo not use this for:\n- Implementation design when the requirement is already clear — use `blueprint`\n- Breaking an approved spec into implementation phases — use `break-spec`\n- Writing code or scaffolding\n\n\u003CHARD-GATE>\nDo NOT design implementation details, invoke `implement`, write code, or scaffold until the product behavior and scope are approved or the user explicitly routes to `blueprint`.\n\u003C\u002FHARD-GATE>\n\nIf the repo declares custom lifecycle hooks, read `..\u002Flifecycle-hooks.md` and honor any phase-boundary hooks before and after spec.\n\nYou may skip steps if the current conversation or `kickoff` already supplied enough context.\n\n## Step 1: Load context\n\nCollect any supplied context:\n- User idea or prompt\n- Ticket title\u002Fbody\u002Fcomments\n- Codebase findings from `kickoff`, if provided\n- Existing plans\u002Fspecs in `plans\u002F`\n- Relevant docs or recent commits, plus the matching `spec_approved` latest pointer entry when the ticket uses custom artifact locations\n\nIf working inside a repo, do light codebase exploration before asking detailed questions. Search for existing patterns, data models, API boundaries, and test coverage that affect the product decision. Record facts, not opinions.\n\nOptional visual routing: only when repo-level `beislid:visual_surfaces` config exists and the effective `spec` mode is active, load the per-skill auxiliary `visual-surface-protocol.md` before the Step 5 approval\u002Frevision surface. It mirrors canonical `.beislid\u002Fvisual-surface-protocol.md` so copied skill installs stay readable. When active, lean toward proposing a supplemental Lavish surface for any non-trivial spec whose requirements, options, flows, data\u002Fstate models, scope boundaries, acceptance outcomes, or decisions can be communicated visually; do not reserve visual routing for UI changes. Use its `BEISLID_VISUAL_PROMPT_V1` envelope for supplemental Lavish HTML review; keep Markdown\u002Fchat spec text canonical, treat freeform visual annotations as advisory, and accept only typed workflow-gate responses that validate\u002Fnormalize to `approve` or `revise`. Unknown, malformed, or freeform-only visual feedback falls back to manual Markdown\u002Fchat review.\n\n## Step 2: Identify unknowns\n\nWrite down 5–10 targeted questions about what you do not know. Frame them as questions, not statements. Focus on:\n- The problem and who feels it\n- Desired behavior\n- Success criteria\n- Scope boundaries\n- Constraints and risks\n- Product decisions that block implementation design\n\n## Step 3: Interview the user\n\nAsk questions one at a time. Prefer multiple choice when possible. If a question can be answered by the codebase or supplied ticket context, answer it from evidence instead of asking.\n\nFor each unresolved question, present:\n- What the codebase\u002Fticket suggests\n- Your recommended answer\n- What decision remains for the user\n\n## Step 4: Propose product directions\n\nWhen there are multiple plausible interpretations or solutions, propose 2–3 product directions with trade-offs and your recommendation. Keep this at the product behavior level. Do not choose file-level implementation details here.\n\nIf the requirement is already obvious and only minor details were clarified, this can be a short paragraph.\n\n## Step 5: Present the spec for approval\n\nBefore writing an artifact, present the proposed spec in concise sections, keep the blocking question out of the draft, and ask for approval or corrections only once in the final response:\n- Problem statement\n- Current state\n- Desired state\n- User stories \u002F acceptance outcomes\n- Key decisions\n- Out of scope\n- Work Contract fields when downstream automation needs a stable handoff\n\nDo not proceed until the user approves the product direction.\n\nOptional Lavish review loop: if Step 1 found active `beislid:visual_surfaces` routing for `spec`, apply `visual-surface-protocol.md` here. Default to proposing the surface for visually-communicable specs; skip only when the spec is trivial, purely linear prose, or the effective mode\u002Ffallback rules prohibit routing. Respect the effective mode exactly: `suggest` only mentions the option; `prompt` asks before generating\u002Fopening and falls back to Markdown\u002Fchat if declined or unattended without permission; `auto` may generate\u002Fopen only within action-policy boundaries and must announce the HTML path before waiting for feedback. The generated surface is supplemental and should use the protocol's `review_spec` prompt, plan\u002Fcomparison layout guidance, and approve\u002Frevise typed gate controls. Normalize typed feedback with the protocol contract or `beislid visual-feedback normalize` semantics; `manual_review`, unknown action\u002Fdecision, malformed payload, freeform-only feedback, or parser-unavailable hosts continue through Markdown\u002Fchat. Copy accepted visual revisions into the Markdown spec, visibly record accepted visual approvals in the canonical Markdown\u002Fchat spec, then ask for or record explicit approval of the canonical Markdown\u002Fchat spec; never let a visual control bypass this approval gate.\n\n## Step 6: Finalize the approved spec\n\nUse this template for the approved spec text. Do NOT include exact file paths or code snippets unless the user explicitly wants an implementation-adjacent note; those belong in `blueprint`.\n\n\u003Cspec-template>\n# [Feature Name]\n\n## Problem Statement\nWhat's broken or missing. Who it affects.\n\n## Current State\nHow things work today. What patterns exist.\n\n## Desired State\nWhat the world looks like after this is released.\n\n## User Stories\n- As a [role], I want [thing] so that [outcome]\n\n## Key Decisions\nDecisions already resolved during the interview, with rationale.\n\n## Out of Scope\nWhat this spec explicitly does not cover.\n\u003C\u002Fspec-template>\n\nThe approved spec text is the source for any artifact. Use the Spec artifact shape from `artifact-templates.md` when writing or summarizing it. Do not add implementation details while writing an artifact.\n\nWhen the next workflow needs stable requirements, finalize a `work-contract-v1` section in or alongside the spec.\n\nRequired structure:\n- Headings: `Kind`, `Status`, `Source`, `Problem`, `Desired Outcome`, `Constraints`, `Acceptance Outcomes`, `Unknowns \u002F Human Decisions`, `Risk Classification`, `Extension Slots`, and `Ownership Boundary`.\n- `Extension Slots` YAML keys: `scope_classification`, `proof_requirements`, `slice_plan`, and `children`.\n\nPopulation rules:\n- `scope_classification` must include `kind`, `confidence`, `rationale`, `recommended_route`, `requires_human_approval`, `requires_split`, and `split_reason`.\n- Valid kinds are `atomic`, `single_pr`, `multi_slice`, `project`, and draft-only `unknown`; valid confidence values are `low`, `medium`, and `high`.\n- Use `atomic` only for bounded, clear, low-branching work; use `single_pr` for one coherent PR; use `multi_slice` for multiple independently shippable slices; use `project` when milestones\u002Fcontracts\u002Fownership boundaries are needed before execution.\n- `unknown` is allowed only for `draft` or `needs-human-decision` contracts and routes to `spec_refinement` with `requires_human_approval: true`.\n- Show the classifier before using it to route downstream work. `requires_human_approval: true` means an extra approval boundary beyond normal spec\u002Fblueprint approval.\n- Require extra approval when classification triggers decomposition, fanout, project planning, contradicts the user's route, or has low-confidence, high-consequence impact. When true, stop, present the classifier and proposed route, wait for explicit approval, record approved\u002Fdeclined, and do not invoke downstream automation until approved. Prefer refinement questions that reduce approval burden over under-classifying.\n- `proof_requirements` uses `proof-requirement-v1`; use `[]` only when no proof has been identified.\n- `slice_plan: null` and `children: []` are reserved for #58.\n- Broad\u002Fproject work should not jump directly to scaffolding by default.\n- Missing product decisions belong under `Unknowns \u002F Human Decisions`; do not invent them to unblock `blueprint`.\n- Work Contracts are Beislið planning semantics, not Rondo execution\u002Fproof\u002Frun state or Memento curated memory.\n\nExamples: typo-level doc fix with no branching → `atomic`; one coherent skill behavior update → `single_pr`; multiple shippable workflow slices → `multi_slice`; broad new-product or cross-system initiative needing milestones\u002Fboundaries → `project`.\n\n## Step 7: Run `spec_approved` lifecycle actions\n\nIf inside a git repo with `.beislid\u002Fworkflow.md`, read only the `beislid:lifecycle_actions` block and execute supported `events.spec_approved.actions[]` entries in order. If no workflow exists, preserve standalone usefulness by offering a local write to `plans\u002F\u003Cfeature-name>-spec.md` after explicit approval. If a workflow exists but no `spec_approved` action is configured, do not ask for ad-hoc tracker\u002Flocal destinations; workflow config controls lifecycle actions.\n\nSupported P0 action shapes:\n\n```yaml\n- name: write-spec-artifact\n  type: artifact\n  approval: prompt # optional; prompt when omitted, auto creates missing target\n  on_failure: prompt # optional; prompt when omitted, or continue | abort\n  path: 'plans\u002F{feature}-spec.md' # optional default\n- name: post-spec-body-to-tracker\n  type: tracker\n  approval: prompt # posts approved spec body through ticket_update.issue_tool \u002F issue_command\n  on_failure: prompt # optional; prompt when omitted, or continue | abort\n- name: run-approved-spec-hook\n  type: cli\n  command: 'planning-hook {event} {ticket_id} {artifact_path}'\n  approval: prompt # required for cli\n  classes: [git-remote] # optional action-policy classes\n  on_failure: prompt # optional; prompt when omitted, or continue | abort\n```\n\nExecute `type: artifact`, `type: tracker`, and `type: cli` under `spec_approved`; skip other providers as reserved. Multiple actions are allowed and run in order. Before each supported action, evaluate action policy: artifact actions use action id `lifecycle.spec_approved.\u003Cname>` with class `workspace-write`; tracker actions use `ticket.update`; CLI actions use `lifecycle.spec_approved.\u003Cname>` with configured `classes` or the conservative default `[workspace-write, git-remote]`. A policy denial records `denied` and skips that action; a policy `ask` boundary must be handled before running.\n\nFor artifact actions, `approval: prompt` asks write\u002Fskip and shows action name, resolved path, and parent directory creation. `approval: auto` writes automatically only when the target does not exist. Existing targets always prompt: overwrite \u002F choose another path \u002F skip. Default path: `plans\u002F{feature}-spec.md`. Supported artifact placeholders are `{feature}`, `{kind}` (`spec`), and `{ticket_id}` when ticket context is known. Derive `{feature}` from the approved spec title, then ticket title, then branch name; ask for a filename stem if none is available. Slug values by lowercasing, replacing non-alphanumeric runs with `-`, collapsing repeats, stripping edge `-`, and keeping names readable (about 60 chars). If `{ticket_id}` is used without ticket context, ask for another path or skip. Paths must be relative, stay inside the repo root (or cwd for standalone fallback), contain no `..`, and end in `.md`. Create parent directories only as part of an approved or auto write.\n\nArtifact content must be the approved spec as primary content. It may add a clearly labeled `## Artifact Context` section with known source event, ticket, branch, and related lifecycle status. Do not alter approved decisions. Treat written spec artifacts as checkpoint-compatible state seeds for fresh-context handoff into `break-spec` or `blueprint`. After the artifact is written, update `.beislid\u002Fcheckpoints\u002Flatest.json` with a `latest.spec_approved` entry containing `event`, `path`, `ticket`, `branch`, `source_skill`, and `written_at` when available; pointer write failures are non-blocking and should be reported.\n\nFor tracker actions, `approval: prompt` asks post\u002Fskip and shows action name plus tracker target. `approval: auto` posts once configured after policy allows it. Tracker actions must use the approved spec body, including the canonical `## Validation\u002FTest Plan` section, as the tracker update body through the configured `ticket_update` issue channel. Treat tracker posts as same-session lifecycle results, not local checkpoint artifacts.\n\nFor CLI actions, `approval` is required. `approval: prompt` asks run\u002Fskip and shows action name, command summary, placeholders used, and action-policy classes. `approval: auto` runs once configured after policy allows it. Supported CLI placeholders are `{ticket_id}`, `{id}` (alias), `{branch}`, `{event}` (`spec_approved`), `{feature}`, `{kind}` (`spec`), and `{artifact_path}` (latest written\u002Fauto-written artifact path for this event, or empty). Pass placeholder values through argv construction when available or shell-quote them before execution; never splice raw branch\u002Fticket text into a shell and never expose the approved spec body as a command-line placeholder.\n\n`on_failure` may be `prompt`, `continue`, or `abort`; omitted means `prompt`. On failed actions, `prompt` asks retry \u002F skip or explicitly override \u002F abort, `continue` warns and routes onward without that side effect, and `abort` stops downstream routing.\n\nRecord lifecycle results as `written`, `auto-written`, `posted`, `ran`, `skipped`, `denied`, `reserved`, `not configured`, or `failed`, with paths\u002Faction names\u002Ftracker targets when available.\n\n## Step 8: Output and route\n\nPrint the approved spec summary, lifecycle status\u002Fpath list, and routing recommendation.\n\nThen route by `scope_classification` when present:\n- `atomic` or `single_pr`: hand off to `blueprint` with the approved spec\u002FWork Contract and any lifecycle artifact path written in this session.\n- `multi_slice`: hand off to `break-spec` with the approved spec\u002FWork Contract and any lifecycle artifact path written in this session.\n- `project`: recommend `spec_refinement` until project boundaries are approved, then hand off to `break-spec`\u002Fslice planning; do not scaffold by default.\n- `unknown`: keep refining; do not hand off to automation as approved.\n- If invoked by `kickoff`, return the approved spec or Work Contract, lifecycle status\u002Fpath, and routing recommendation to `kickoff`.\n","# Spec\n\nShape an idea, vague ticket, or open product question into a lightweight product spec. This is the product\u002Frequirements gate, not the implementation-design gate.\n\nUse this for:\n- Open-ended brainstorming\n- Vague tickets that are not ready for implementation design\n- Product behavior, scope, and acceptance criteria discovery\n- Writing a lightweight spec\u002FPRD before `blueprint`\n\nDo not use this for:\n- Implementation design when the requirement is already clear — use `blueprint`\n- Breaking an approved spec into implementation phases — use `break-spec`\n- Writing code or scaffolding\n\n\u003CHARD-GATE>\nDo NOT design implementation details, invoke `implement`, write code, or scaffold until the product behavior and scope are approved or the user explicitly routes to `blueprint`.\n\u003C\u002FHARD-GATE>\n\nIf the repo declares custom lifecycle hooks, read `..\u002Flifecycle-hooks.md` and honor any phase-boundary hooks before and after spec.\n\nYou may skip steps if the current conversation or `kickoff` already supplied enough context.\n\n## Step 1: Load context\n\nCollect any supplied context:\n- User idea or prompt\n- Ticket title\u002Fbody\u002Fcomments\n- Codebase findings from `kickoff`, if provided\n- Existing plans\u002Fspecs in `plans\u002F`\n- Relevant docs or recent commits, plus the matching `spec_approved` latest pointer entry when the ticket uses custom artifact locations\n\nIf working inside a repo, do light codebase exploration before asking detailed questions. Search for existing patterns, data models, API boundaries, and test coverage that affect the product decision. Record facts, not opinions.\n\nOptional visual routing: only when repo-level `beislid:visual_surfaces` config exists and the effective `spec` mode is active, load the per-skill auxiliary `visual-surface-protocol.md` before the Step 5 approval\u002Frevision surface. It mirrors canonical `.beislid\u002Fvisual-surface-protocol.md` so copied skill installs stay readable. When active, lean toward proposing a supplemental Lavish surface for any non-trivial spec whose requirements, options, flows, data\u002Fstate models, scope boundaries, acceptance outcomes, or decisions can be communicated visually; do not reserve visual routing for UI changes. Use its `BEISLID_VISUAL_PROMPT_V1` envelope for supplemental Lavish HTML review; keep Markdown\u002Fchat spec text canonical, treat freeform visual annotations as advisory, and accept only typed workflow-gate responses that validate\u002Fnormalize to `approve` or `revise`. Unknown, malformed, or freeform-only visual feedback falls back to manual Markdown\u002Fchat review.\n\n## Step 2: Identify unknowns\n\nWrite down 5–10 targeted questions about what you do not know. Frame them as questions, not statements. Focus on:\n- The problem and who feels it\n- Desired behavior\n- Success criteria\n- Scope boundaries\n- Constraints and risks\n- Product decisions that block implementation design\n\n## Step 3: Interview the user\n\nAsk questions one at a time. Prefer multiple choice when possible. If a question can be answered by the codebase or supplied ticket context, answer it from evidence instead of asking.\n\nFor each unresolved question, present:\n- What the codebase\u002Fticket suggests\n- Your recommended answer\n- What decision remains for the user\n\n## Step 4: Propose product directions\n\nWhen there are multiple plausible interpretations or solutions, propose 2–3 product directions with trade-offs and your recommendation. Keep this at the product behavior level. Do not choose file-level implementation details here.\n\nIf the requirement is already obvious and only minor details were clarified, this can be a short paragraph.\n\n## Step 5: Present the spec for approval\n\nBefore writing an artifact, present the proposed spec in concise sections, keep the blocking question out of the draft, and ask for approval or corrections only once in the final response:\n- Problem statement\n- Current state\n- Desired state\n- User stories \u002F acceptance outcomes\n- Key decisions\n- Out of scope\n- Work Contract fields when downstream automation needs a stable handoff\n\nDo not proceed until the user approves the product direction.\n\nOptional Lavish review loop: if Step 1 found active `beislid:visual_surfaces` routing for `spec`, apply `visual-surface-protocol.md` here. Default to proposing the surface for visually-communicable specs; skip only when the spec is trivial, purely linear prose, or the effective mode\u002Ffallback rules prohibit routing. Respect the effective mode exactly: `suggest` only mentions the option; `prompt` asks before generating\u002Fopening and falls back to Markdown\u002Fchat if declined or unattended without permission; `auto` may generate\u002Fopen only within action-policy boundaries and must announce the HTML path before waiting for feedback. The generated surface is supplemental and should use the protocol's `review_spec` prompt, plan\u002Fcomparison layout guidance, and approve\u002Frevise typed gate controls. Normalize typed feedback with the protocol contract or `beislid visual-feedback normalize` semantics; `manual_review`, unknown action\u002Fdecision, malformed payload, freeform-only feedback, or parser-unavailable hosts continue through Markdown\u002Fchat. Copy accepted visual revisions into the Markdown spec, visibly record accepted visual approvals in the canonical Markdown\u002Fchat spec, then ask for or record explicit approval of the canonical Markdown\u002Fchat spec; never let a visual control bypass this approval gate.\n\n## Step 6: Finalize the approved spec\n\nUse this template for the approved spec text. Do NOT include exact file paths or code snippets unless the user explicitly wants an implementation-adjacent note; those belong in `blueprint`.\n\n\u003Cspec-template>\n# [Feature Name]\n\n## Problem Statement\nWhat's broken or missing. Who it affects.\n\n## Current State\nHow things work today. What patterns exist.\n\n## Desired State\nWhat the world looks like after this is released.\n\n## User Stories\n- As a [role], I want [thing] so that [outcome]\n\n## Key Decisions\nDecisions already resolved during the interview, with rationale.\n\n## Out of Scope\nWhat this spec explicitly does not cover.\n\u003C\u002Fspec-template>\n\nThe approved spec text is the source for any artifact. Use the Spec artifact shape from `artifact-templates.md` when writing or summarizing it. Do not add implementation details while writing an artifact.\n\nWhen the next workflow needs stable requirements, finalize a `work-contract-v1` section in or alongside the spec.\n\nRequired structure:\n- Headings: `Kind`, `Status`, `Source`, `Problem`, `Desired Outcome`, `Constraints`, `Acceptance Outcomes`, `Unknowns \u002F Human Decisions`, `Risk Classification`, `Extension Slots`, and `Ownership Boundary`.\n- `Extension Slots` YAML keys: `scope_classification`, `proof_requirements`, `slice_plan`, and `children`.\n\nPopulation rules:\n- `scope_classification` must include `kind`, `confidence`, `rationale`, `recommended_route`, `requires_human_approval`, `requires_split`, and `split_reason`.\n- Valid kinds are `atomic`, `single_pr`, `multi_slice`, `project`, and draft-only `unknown`; valid confidence values are `low`, `medium`, and `high`.\n- Use `atomic` only for bounded, clear, low-branching work; use `single_pr` for one coherent PR; use `multi_slice` for multiple independently shippable slices; use `project` when milestones\u002Fcontracts\u002Fownership boundaries are needed before execution.\n- `unknown` is allowed only for `draft` or `needs-human-decision` contracts and routes to `spec_refinement` with `requires_human_approval: true`.\n- Show the classifier before using it to route downstream work. `requires_human_approval: true` means an extra approval boundary beyond normal spec\u002Fblueprint approval.\n- Require extra approval when classification triggers decomposition, fanout, project planning, contradicts the user's route, or has low-confidence, high-consequence impact. When true, stop, present the classifier and proposed route, wait for explicit approval, record approved\u002Fdeclined, and do not invoke downstream automation until approved. Prefer refinement questions that reduce approval burden over under-classifying.\n- `proof_requirements` uses `proof-requirement-v1`; use `[]` only when no proof has been identified.\n- `slice_plan: null` and `children: []` are reserved for #58.\n- Broad\u002Fproject work should not jump directly to scaffolding by default.\n- Missing product decisions belong under `Unknowns \u002F Human Decisions`; do not invent them to unblock `blueprint`.\n- Work Contracts are Beislið planning semantics, not Rondo execution\u002Fproof\u002Frun state or Memento curated memory.\n\nExamples: typo-level doc fix with no branching → `atomic`; one coherent skill behavior update → `single_pr`; multiple shippable workflow slices → `multi_slice`; broad new-product or cross-system initiative needing milestones\u002Fboundaries → `project`.\n\n## Step 7: Run `spec_approved` lifecycle actions\n\nIf inside a git repo with `.beislid\u002Fworkflow.md`, read only the `beislid:lifecycle_actions` block and execute supported `events.spec_approved.actions[]` entries in order. If no workflow exists, preserve standalone usefulness by offering a local write to `plans\u002F\u003Cfeature-name>-spec.md` after explicit approval. If a workflow exists but no `spec_approved` action is configured, do not ask for ad-hoc tracker\u002Flocal destinations; workflow config controls lifecycle actions.\n\nSupported P0 action shapes:\n\n```yaml\n- name: write-spec-artifact\n  type: artifact\n  approval: prompt # optional; prompt when omitted, auto creates missing target\n  on_failure: prompt # optional; prompt when omitted, or continue | abort\n  path: 'plans\u002F{feature}-spec.md' # optional default\n- name: post-spec-body-to-tracker\n  type: tracker\n  approval: prompt # posts approved spec body through ticket_update.issue_tool \u002F issue_command\n  on_failure: prompt # optional; prompt when omitted, or continue | abort\n- name: run-approved-spec-hook\n  type: cli\n  command: 'planning-hook {event} {ticket_id} {artifact_path}'\n  approval: prompt # required for cli\n  classes: [git-remote] # optional action-policy classes\n  on_failure: prompt # optional; prompt when omitted, or continue | abort\n```\n\nExecute `type: artifact`, `type: tracker`, and `type: cli` under `spec_approved`; skip other providers as reserved. Multiple actions are allowed and run in order. Before each supported action, evaluate action policy: artifact actions use action id `lifecycle.spec_approved.\u003Cname>` with class `workspace-write`; tracker actions use `ticket.update`; CLI actions use `lifecycle.spec_approved.\u003Cname>` with configured `classes` or the conservative default `[workspace-write, git-remote]`. A policy denial records `denied` and skips that action; a policy `ask` boundary must be handled before running.\n\nFor artifact actions, `approval: prompt` asks write\u002Fskip and shows action name, resolved path, and parent directory creation. `approval: auto` writes automatically only when the target does not exist. Existing targets always prompt: overwrite \u002F choose another path \u002F skip. Default path: `plans\u002F{feature}-spec.md`. Supported artifact placeholders are `{feature}`, `{kind}` (`spec`), and `{ticket_id}` when ticket context is known. Derive `{feature}` from the approved spec title, then ticket title, then branch name; ask for a filename stem if none is available. Slug values by lowercasing, replacing non-alphanumeric runs with `-`, collapsing repeats, stripping edge `-`, and keeping names readable (about 60 chars). If `{ticket_id}` is used without ticket context, ask for another path or skip. Paths must be relative, stay inside the repo root (or cwd for standalone fallback), contain no `..`, and end in `.md`. Create parent directories only as part of an approved or auto write.\n\nArtifact content must be the approved spec as primary content. It may add a clearly labeled `## Artifact Context` section with known source event, ticket, branch, and related lifecycle status. Do not alter approved decisions. Treat written spec artifacts as checkpoint-compatible state seeds for fresh-context handoff into `break-spec` or `blueprint`. After the artifact is written, update `.beislid\u002Fcheckpoints\u002Flatest.json` with a `latest.spec_approved` entry containing `event`, `path`, `ticket`, `branch`, `source_skill`, and `written_at` when available; pointer write failures are non-blocking and should be reported.\n\nFor tracker actions, `approval: prompt` asks post\u002Fskip and shows action name plus tracker target. `approval: auto` posts once configured after policy allows it. Tracker actions must use the approved spec body, including the canonical `## Validation\u002FTest Plan` section, as the tracker update body through the configured `ticket_update` issue channel. Treat tracker posts as same-session lifecycle results, not local checkpoint artifacts.\n\nFor CLI actions, `approval` is required. `approval: prompt` asks run\u002Fskip and shows action name, command summary, placeholders used, and action-policy classes. `approval: auto` runs once configured after policy allows it. Supported CLI placeholders are `{ticket_id}`, `{id}` (alias), `{branch}`, `{event}` (`spec_approved`), `{feature}`, `{kind}` (`spec`), and `{artifact_path}` (latest written\u002Fauto-written artifact path for this event, or empty). Pass placeholder values through argv construction when available or shell-quote them before execution; never splice raw branch\u002Fticket text into a shell and never expose the approved spec body as a command-line placeholder.\n\n`on_failure` may be `prompt`, `continue`, or `abort`; omitted means `prompt`. On failed actions, `prompt` asks retry \u002F skip or explicitly override \u002F abort, `continue` warns and routes onward without that side effect, and `abort` stops downstream routing.\n\nRecord lifecycle results as `written`, `auto-written`, `posted`, `ran`, `skipped`, `denied`, `reserved`, `not configured`, or `failed`, with paths\u002Faction names\u002Ftracker targets when available.\n\n## Step 8: Output and route\n\nPrint the approved spec summary, lifecycle status\u002Fpath list, and routing recommendation.\n\nThen route by `scope_classification` when present:\n- `atomic` or `single_pr`: hand off to `blueprint` with the approved spec\u002FWork Contract and any lifecycle artifact path written in this session.\n- `multi_slice`: hand off to `break-spec` with the approved spec\u002FWork Contract and any lifecycle artifact path written in this session.\n- `project`: recommend `spec_refinement` until project boundaries are approved, then hand off to `break-spec`\u002Fslice planning; do not scaffold by default.\n- `unknown`: keep refining; do not hand off to automation as approved.\n- If invoked by `kickoff`, return the approved spec or Work Contract, lifecycle status\u002Fpath, and routing recommendation to `kickoff`.\n",{"name":5,"description":7},"main","bba9b487e6731261bcfe35e17ea03269cf8df0a1","resources\u002Fbeislid\u002Fskills\u002Fspec","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal","https:\u002F\u002Fnopal.sh","https:\u002F\u002Fgithub.com\u002Fsandsower","vic.valenzuela.is","Product-minded software engineer","2026-10-15 00:33:24",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add sandsower\u002Fnopal --skill spec",{"tool":39,"command":40},"bunx","bunx skills add sandsower\u002Fnopal --skill spec",{"tool":42,"command":43},"pnpm","pnpm dlx skills add sandsower\u002Fnopal --skill spec","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fspec\u002FSKILL.md",true,[47,53],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"artifact-templates.md","REFERENCE","参考文档",8423,"https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fspec\u002Fartifact-templates.md",{"path":54,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":55,"sourceHtmlUrl":56},"visual-surface-protocol.md",26191,"https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fspec\u002Fvisual-surface-protocol.md",[58,65,72,79,86,93],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":19},"01M78GRB8RYKHHGYZF3FGN9Y5P","walk-the-diff","Use when the user wants to walk through their local changes interactively before or instead of a PR review — e.g., \"guided review\", \"walk me through the changes\", \"review my branch\", \"let's go through the diff\". Also use when the user explicitly invokes the walk-the-diff skill through the host agent. This is for interactive human-facing walkthroughs of code changes, not automated code review.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fwalk-the-diff\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M78GR8YWH8JMB2E9RRSAR7SF","verify","Use before claiming work is complete, fixed, or passing. Requires running verification commands and reading output before making any success claims. Evidence before assertions.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fverify\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":19},"01M78GQXP7EJ8WHHZJ4H1NNN0N","show-me","Use when the user asks for a visual HTML evidence\u002Fexplanation artifact: 'show-me', 'show me this', 'make a proof deck', 'make an HTML deck', 'visualize this as HTML', 'this would be better as slides', 'make this visual', 'create a walkthrough deck', 'document this visually', or 'review this visually'. Manual\u002Fuser-requested only in v1. Produces a polished local HTML portfolio\u002Fdeck with evidence, media, logs, provenance, and clear missing-evidence markers.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fshow-me\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M78GQS6VPCC70BNE86HSAF6S","setup","Configure Beislið project config (`.beislid\u002Fworkflow.md`) interactively, or update an installed Beislið distribution. Use when the user says \"setup\", \"\u002Fsetup\", \"setup update\", \"\u002Fsetup update\", \"update beislid\", \"configure beislid\", \"set up workflow\", \"add scopes\", \"change ticket source\", \"reconfigure beislid\", or any intent to add, change, or remove a workflow.md section. Walks through sections conversationally; shows a diff before any destructive write.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fsetup\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":19},"01M78GQNPYS736XNYDGKW5N7F4","rinse","Use when the user asks to 'rinse', 'run a review\u002Ffix loop', 'iterate on review findings', 'review and fix until clean', or harden a branch before PR handoff. Orchestrates side-effect-free review, user-approved fixes, verification, reruns, and redesign stop conditions.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Frinse\u002FSKILL.md",[],[],{"skillNo":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":19},"01M78GQJE55XWGK0N17FC3BKEP","review","Review a local diff or supplied patch against requirements. Produces severity-categorized findings and a readiness verdict. Side-effect-free primitive used by ready-for-review, future rinse-style loops, PR patrol workflows, and ad-hoc review requests.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Freview\u002FSKILL.md",[],[]]