atelier-fashion/adlc-toolkit/tasks
Claude Code and agent skill by atelier-fashion/adlc-toolkit
Trust tier 3
- community
Source check pending
Checking that the SKILL.md file is still in the repository.
Install
npx vskill@latest install atelier-fashion/adlc-toolkit/tasksSKILL.md
Description
Create manifest/SKILL.md — a standalone, read-only skill that derives a cross-session view of in-flight ADLC work from the remote (open PRs + pushed feat/REQ-* branches), enriches each entry with component/domain, computes a coarse overlap report, and renders an advisory table. This is the foundational artifact; the wiring tasks (066–068) depend on it.
Files to Create/Modify
manifest/SKILL.md— the new skill (create)
Acceptance Criteria
- [ ] Skill file follows the canonical skeleton: frontmatter (
name: manifest,description,argument-hint),# /manifest — …title,## Ethosinjection macro (!sh .adlc/partials/ethos-include.sh 2>/dev/null || sh ~/.claude/skills/partials/ethos-include.sh`),## Context,## Input(optional REQ filter /self),## Prerequisites(git repo + reachableorigin; fail clearly otherwise),## Instructions,## Quality Checklist`. - [ ] Enumerates in-flight REQs from BOTH
gh pr list --state open --json number,headRefName,author,isDraft,createdAt,url --limit 200(incl. drafts) AND remotefeat/REQ-*branches (git branch -r --list 'origin/feat/REQ-*'), deduped by REQ id (PR entry wins; branch-only added when no PR). (BR-3, BR-4) - [ ] Maps branch → REQ via
feat/REQ-<digits>-<slug>; non-matching branches are silently ignored. (BR-4) - [ ] Every remote-derived identifier is sanitized before shell use: REQ ids validated against
^REQ-[0-9]{3,6}$; substituted branch/path values single-quoted; any path segment equal to..rejected beforegit show. (BR-5, LESSON-008) - [ ] Component/domain enrichment tries local frontmatter first, then
git show origin/<branch>:.adlc/specs/<REQ>-*/requirement.md, thenunknown; an entry is never dropped on enrichment failure. (BR-11, BR-12) - [ ] Coarse overlap report flags pairs sharing
componentORdomain(labels which), presented as ADVISORY only with an explicit "no action enforced" note. (BR-8) - [ ] Includes the current session's own REQ when one is in context, marked as self. (BR-13)
- [ ] Graceful degradation: with
ghunavailable/unauthenticated, enumerate from branches only, annotate that PR data is missing, exit 0 — never hard-fail. (BR-6) - [ ] Strictly read-only: no branch/PR/file/state mutation;
git statusclean after a run. (BR-2) No stored manifest file is created. (BR-1) - [ ] Network cost is O(1) in the number of in-flight REQs: reuse caller's fetch in pre-flight contexts / one own
git fetchstandalone, one batchedgh pr list, localgit showenrichment — no per-branch API calls. (BR-14) - [ ] Output table columns:
REQ | author | branch/PR | state | component/domain | opened-at(+ self marker). (AC from requirement) - [ ]
python3 tools/lint-skills/check.pypasses clean (no new findings). - [ ] Dogfood: running
/manifestin this repo listsfeat/REQ-482-manifest-skillas in-flight (self) and exits 0.
Technical Notes
- Mirror
status/SKILL.mdfor structure and table style; mirrorproceed/SKILL.md/template-drift/SKILL.mdfor the identifier-sanitization + path-traversal-rejection patterns. - No Kimi (ADR-3): do NOT reference
ADLC_DISABLE_KIMI,kimi-gate.sh, or telemetry — this skill does git/gh I/O, not bulk reads, and referencing the gate would trip the lint canonical-helper check. - POSIX/BSD bash only (conventions "Bash in skills"):
grep -oEnot-oP; nolocalinside`sh/`shellfences (use`bashiflocalis truly needed); keep any shell function defined-and-invoked within the same fenced block (cross-fence-fn lint check — LESSON-020). - Keep
$(/)and$((/))balanced within each fence (lint shell-balance check). statevalues:draft(PR isDraft) |ready(open non-draft PR) |no-pr(branch only).
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryChecking.adlc/specs/REQ-482-manifest-skill/tasks/TASK-065-create-manifest-skill.md
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/atelier-fashion/adlc-toolkit/tasks)