← Back to list

agent-instruction-maintenance
by pevd950
⭐ 0🍴 0📅 Jan 25, 2026
SKILL.md
name: agent-instruction-maintenance description: "Review, update, and synchronize agent instructions and configs across Codex and Claude (and optionally VS Code/Copilot); use when creating or refactoring agent prompts, skills, or chatmodes (trigger keywords: agent instructions, prompt review, skill update, chatmode, subagent)."
Agent Instruction Maintenance
What this skill does
- Audits agent instructions for clarity, correctness, and consistency.
- Validates frontmatter and minimal schemas for detected ecosystems.
- Aligns tool-specific variants while keeping behavior consistent.
When to use it
- Trigger phrases: "agent instructions", "prompt review", "skill update", "chatmode", "subagent".
- Use after creating or editing instructions, or when standards change.
Step-by-step workflow
- Discover which ecosystems exist in the repo or user config.
- Load local guidance (AGENTS.md, CLAUDE.md, copilot-instructions) when present.
- Validate minimal frontmatter for each file type (see references).
- Review content for clarity, scope, and actionability; remove ambiguity.
- Ensure intent parity across tools while avoiding contradictions.
- Keep instructions lean; move deep details into references.
- Propose updates with before/after snippets and rationale.
Default prioritization
- Prefer Codex-first guidance unless the user asks for tool-specific changes.
- Prefer Skills (SKILL.md) and AGENTS.md for portable, shared behavior.
- Apply tool-specific files only when a feature is unique to that tool.
Tool-specific references
- Codex:
references/codex.md - Claude Opus:
references/claude-code.md - VS Code/Copilot (optional):
references/copilot-vscode.md - Tool choice (commands vs skills vs agents):
references/tool-selection.md
Expected outputs / formatting
- Summary of findings and risk areas.
- Required fixes (schema/format) vs recommended improvements (clarity/consistency).
- Proposed edits with brief rationale.
- Updated files when approved.
Example prompts
- "Review our skills and normalize naming/trigger phrases."
- "Validate agent frontmatter across Claude and Copilot files."
- "Update Codex skills to match new project standards."
Score
Total Score
50/100
Based on repository quality metrics
✓SKILL.md
SKILL.mdファイルが含まれている
+20
○LICENSE
ライセンスが設定されている
0/10
○説明文
100文字以上の説明がある
0/10
○人気
GitHub Stars 100以上
0/15
○最近の活動
3ヶ月以内に更新がある
0/10
○フォーク
10回以上フォークされている
0/5
✓Issue管理
オープンIssueが50未満
+5
✓言語
プログラミング言語が設定されている
+5
○タグ
1つ以上のタグが設定されている
0/5
Reviews
💬
Reviews coming soon