Audit a Claude Code install
Find the Claude Code failures that produce no error message: dead hooks, rejected skills, missing plugins, oversized memory files.
Free
Works with the AI Agents you already use
Audit a Claude Code install
Example session with this skill installed
My PostToolUse hook stopped running after I renamed a project directory, and one of my skills never triggers. Nothing shows an error.
- Read your context and instructions
- Compiled the audit a claude
Summary
8 checks run over ~/.claude. Nothing was created, modified or deleted.
| Check | Verdict | Evidence |
|---|---|---|
| settingsParse | ok | valid JSON |
| inlineSecret | ok | no plaintext credential |
| apiKeyHelper | ok | command resolves |
| hooks | warn | PostToolUse -> ~/.claude/hooks/fmt.sh (file missing) |
| skills | warn | frontmatter name "reviewer" != directory "code-review" |
| memoryFile | warn | CLAUDE.md is 19.8k tokens, injected every turn |
| plugins | ok | 4 registered, 4 loadable |
| backups | warn | 2 stale settings copies, each holds the same key |
What to fix first
- The PostToolUse hook has not run since the rename — restore or re-point that script.
- A skill whose frontmatter name differs from its directory never fires. Rename one to match the other.
- CLAUDE.md is a per-turn tax, not a one-time cost. Trim or split it.
Read-only. Nothing was created, modified or deleted.
Connects securely to your tools. The creator never sees your data.
About this skill
Some Claude Code failures are silent. A hook points at a file that no longer exists. A skill's frontmatter no longer parses, so it is quietly skipped. A plugin is installed but nothing loads it. A CLAUDE.md grew to 20k tokens and is now paid on every single turn.
None of these produce an error. All of them are checkable.
This skill runs eight checks over the config directory - settings, hooks, skills, memory files, plugins, backups, transcripts, and credential-shaped values in settings keys that name one - and reports each as ok, warn, or unmeasured.
Unmeasured is a distinct state, not a pass. A check that could not read anything says so, because "I could not look" and "I looked and it is fine" are different answers.
Read-only. Node 20+.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.
Read-only access only. Requires Node.js >= 20. Cannot modify settings.json or delete backup files. Limited to the ~/.claude directory.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 2 days ago
- Free with an Agensi account
- Software
Node.js