The missing linter and LSP for AI-coding-assistant configs (CLAUDE.md, SKILL.md, hooks, MCP).
Agnix validates agent configuration files — CLAUDE.md, AGENTS.md, SKILL.md frontmatter, hook scripts, MCP settings — with 156 rules and autofixes. Plugins for VS Code, Zed, Neovim. Treat it like ESLint for your agent setup.
claude mcp add agnix-skill -- git clone https://github.com/avifenesh/agnix ~/.claude/skills/agnix
ワンライナー。claude mcp list で確認、claude mcp remove で削除。
ユースケース
実用的な使い方: Agnix (Agent Config Linter)
Enforce skill quality in CI before merging to the team's shared skill repo
👤 Teams maintaining a skill library⏱ ~15 minintermediate
使うタイミング: You don't want a broken SKILL.md frontmatter to ship and break everyone's agent.
前提条件
Server/skill installed and authenticated — See repo README
フロー
Set up CI
Add agnix check to GitHub Actions on every PR touching .claude/ or skills/.✓ コピーしました
→ Workflow file added
Fix locally
Run agnix fix on the staged files. Show me the diff before committing.✓ コピーしました
→ Autofix diff
結果: No more bad-frontmatter incidents in production.
注意点
Some rules are opinion-shaped (e.g., max description length). Tune in .agnix.toml rather than ignoring globally. — Some rules are opinion-shaped (e.g., max description length). Tune in .agnix.toml rather than ignoring globally.
コストと制限
運用コスト
APIクォータ
See provider docs for rate limits
呼び出しあたりのトークン
Varies by tool
金額
See repo README for pricing details
ヒント
Cache tool results and avoid repeated identical calls.
セキュリティ
権限、シークレット、影響範囲
認証情報の保管: Use environment variables; never commit secrets
データ送信先: Tool calls go to the provider's API as documented
トラブルシューティング
よくあるエラーと対処法
rule too strict
Tune in .agnix.toml — every rule has severity/disable settings.
確認: agnix check --explain <rule-id>
LSP doesn't activate
Make sure the editor extension matches your agnix CLI version.