agent-architecture-audit

Jump to install ↓

エージェントおよび LLM アプリケーション向けのフルスタック診断。12 層のエージェントスタックにおけるラッパーリグレッション、メモリ汚染、ツール規律の失敗、隠れた修復ループ、レンダリング破損を監査します。重要度順の発見事項とコードファー…

by: affaan-m
Stars: ★ 243,286
License: MIT

Where it comes from

We index this skill; we do not host it. The app downloads it straight from the repository below, so the author keeps their credit and their licence.

Repository: affaan-m/everything-claude-code
Branch: main
Path: docs/ja-JP/skills/agent-architecture-audit
Written in: Japanese
Stars: ★ 243,286
License: MIT

Read it on GitHub ↗

Install it

Read it first, then install. A skill is code that runs inside your agent.

Into

Paste this into Claude, Codex, or whichever agent you use. It reads the source first and tells you what it found before installing.

Read the SKILL.md and its companion files at https://github.com/affaan-m/everything-claude-code/tree/main/docs/ja-JP/skills/agent-architecture-audit first.

Then tell me three things:
1. What this skill actually does
2. Whether it runs shell commands, reaches the network, writes files, or touches credentials
3. Whether it is worth adding to this project

If it looks fine, install it: npx skills add affaan-m/everything-claude-code --skill agent-architecture-audit -a claude-code

Read it on GitHub ↗

Where it lands

Claude Code: .claude/skills/<name>/
Claude: Project knowledge
Codex: AGENTS.md
Cursor: .cursor/rules/<name>.mdc
ChatGPT: Custom instructions

Add to a tree

Something wrong with this skill?

Report it in the app and it drops out of the catalogue right away.

agent-architecture-audit

Where it comes from

More from this repository