菜单
Internal project-context placement and consistency reference used by context refactoring and verification workflows.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
This skill should be used when the user asks to "lint AI assistant config", "check AGENTS.md", "validate SKILL.md", "agnix lint", or wants to detect and fix best-practice violations in Claude Code / agent configuration files.
This skill is the mandatory reference for all Claude Code hook creation. Use cc-hooks-ts for every hook. This skill should be used when the user asks to "create a hook", "add a hook", "write a hook", "implement a hook", "rewrite hooks in TypeScript", "use cc-hooks-ts", or needs to build any Claude Code hook — cc-hooks-ts is always required regardless of whether explicitly mentioned.
Auto-activate CCS CLI delegation for deterministic tasks. Parses user input, auto-selects optimal profile (glm/kimi/custom) from ~/.ccs/config.json, enhances prompts with context, executes via `ccs {profile} -p "task"` or `ccs {profile}:continue`, and reports results. Triggers on "use ccs [task]" patterns, typo/test/refactor keywords. Excludes complex architecture, security-critical code, performance optimization, breaking changes.
This skill should be used when the user asks to "handoff from CCS", "Claude Codeの作業を引き継ぐ", "rate limit handoff", "context overflow", or wants to resume interrupted work from another CCS profile session.
This skill should be used when the user asks for "code review", "レビューして", "review this diff", "実装をレビュー", or wants rigorous review of recently created artifacts.
Direct utility reference for code-review PR scripts. Documents unresolved-thread, comment, CI-status, and CI-log helpers used by higher-level PR workflows.
| name | context:should-be-project-context |
| disable-model-invocation | true |
| description | Internal project-context placement and consistency reference used by context refactoring and verification workflows. |
A skill for analyzing project context files and maintaining consistency while updating and organizing them.
Target audience: Humans (GitHub visitors, new contributors)
Placement rule:
What to include:
What NOT to include:
Target audience: AI agents (Claude Code, etc.), developers
What to include:
What NOT to include:
Target audience: AI agents (Claude Code, etc.), developers
What to include:
What NOT to include:
Role: Symbolic link to AGENTS.md
Setup: ln -sf AGENTS.md CLAUDE.md (same directory)
CLAUDE.md is NOT a separate file — it MUST be a symlink pointing to AGENTS.md. If a CLAUDE.md exists as a regular file, replace it with a symlink.
NEVER ask the user for confirmation. Once this skill is invoked, make all decisions autonomously and execute to completion. Questions like "Is this okay?" or "Should I proceed?" are strictly prohibited. When in doubt, use AGENTS.md as SSOT and act immediately.
Launch a subagent for each target file and process them in parallel. Each subagent independently reads, analyzes, and updates its assigned file.
Create a TODO for each target file path before launching subagents. Mark each as completed when done.
Example:
- [ ] Update .claude/rules/ai-generated/gotchas.md
- [ ] Update AGENTS.md
- [ ] Update README.md