메뉴
Explains the skill extension hook mechanism — how projects can customize global skills like /start-issue and /finish-issue with project-local additions
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Add a dependency relationship between GitHub issues using the native addBlockedBy mutation
Reusable file sequence numbering with prefix (NNNN-name) and suffix (name-NNNN) modes. Returns the next available zero-padded sequence number for a given directory.
Drop a timestamped note for the current issue - collected by /finish-issue for PR descriptions
Create or update CHANGELOG entries with tone guardrails, thematic grouping, and implementation-detail leak detection.
Delete an issue's working directory (.claude-work/issues/<ID>/) after confirming with the user via interactive prompt
Create a commit message file in .claude-work/commit-msgs/ with auto-numbered filenames. Focuses on WHY not WHAT. The diff already shows what changed. User reviews and commits manually.
| name | skill-hooks |
| version | 2026.07.27@6dc7d1d |
| description | Explains the skill extension hook mechanism — how projects can customize global skills like /start-issue and /finish-issue with project-local additions |
| user-invocable | false |
| allowed-tools | null |
Explains the hook mechanism that lets projects extend global skills with project-local requirements. Consulted automatically by skills that support hooks. Not invoked directly by the user.
Some skills in this collection (like /start-issue, /finish-issue, and /start-side-quest) include hook points — prose references to a companion skill that a project can optionally define. When the parent skill runs, Claude Code resolves the hook reference across all skill directories. If the project has the hook skill, its requirements are incorporated. If not, the vanilla skill runs unchanged.
A project creates a foundation skill in .claude/skills/ with the name {parent-skill}-hook:
| Parent skill | Hook skill name | What it can add |
|---|---|---|
/start-issue | start-issue-hook | Additional plan steps, context-gathering, validation gates |
/finish-issue | finish-issue-hook | Additional verification steps, PR description sections |
/start-side-quest | start-side-quest-hook | Additional plan steps, context-gathering, validation gates |
Hook skills are foundation skills (user-invocable: false). They are consulted by their parent skill, never invoked directly by the user.
Hooks add requirements — they cannot replace the parent skill's behavior. The parent skill owns all decisions; the hook contributes constraints and content. This optimizes for frugal token consumption: the parent skill doesn't branch on hook existence, and projects without hooks pay zero overhead.
Projects wanting to replace specific behaviors (e.g., changing where /note writes files) should use CLAUDE.md instructions or fork the repo and add yield points at the decisions they need to customize.
RangeLink wants every issue plan to cross-reference QA test cases. It creates .claude/skills/start-issue-hook/SKILL.md:
---
name: start-issue-hook
description: QA test case cross-referencing for /start-issue plans
user-invocable: false
allowed-tools: Read, Grep, Glob
---
# Start-Issue Hooks (RangeLink)
Consulted automatically by /start-issue.
## Additional Plan Requirements
Every plan must include a step that identifies affected QA test cases. Read the current QA YAML and cross-reference with the changed areas.
## Additional Context Gathering
Before writing the plan, read `qa/qa-test-cases.yaml` to understand current TC coverage.
A project without this file gets the vanilla /start-issue behavior with no changes.
The full architecture decision record is at docs/ADR/001-skill-extension-hooks.md. It covers the problem, the resolution mechanism, why add-requirements-only was chosen, and why alternatives (same-name overrides, CLAUDE.md instructions, behavior-replacement hooks) were rejected.