audit-config — Claude Code command: config audit, how to install
audit-config counts what sits in your Claude config per surface and prints one compact table with a short verdict on where things have grown.
MIT ★ 1,198 updated 2026-10-23
/plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor
/plugin install ccam-config@claude-code-agent-monitor-pluginsWhat audit-config does
The command calls curl -s http://localhost:4820/api/cc-config/overview on the Config Explorer in Agent Monitor and shapes the JSON into a one-screen table. Rows cover skills, agents, commands, output styles, MCP servers and hooks; columns are user, project and total. Plugins and memory appear as a single figure, with enabled and disabled plugins split out.
Above the table it echoes roots.claudeHome and roots.projectRoot, so you can tell which directories were inspected. One closing line names the biggest surface and points at sprawl. Every figure comes from the API response.
The dashboard has to be up at http://localhost:4820 first. When it isn't, the command says so and points you to npm start from the repo root. On Codex the same prompt lives at ~/.codex/prompts/audit-config.md.
What it helps with
- count skills and subagents split between user and project scope
- spot heavy user-scope commands and skills before cleaning up
~/.claude - check enabled versus disabled plugins alongside marketplace count
- see which MCP servers and hooks are wired into the current project
- capture a config snapshot before changing settings
Example prompts for the agent
- Run
/audit-configand show me the table for this repo. - Audit my Claude config: how many skills and hooks sit in user versus project?
- Run
/audit-configand tell me which surface is the largest.
How to install audit-config
1Plugin marketplace
/plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor/plugin install ccam-config@claude-code-agent-monitor-plugins2Manually: put the files in place
.claude/commands/audit-config.mdCall it with /audit-config. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/audit-config.mdCall it with /prompts:audit-config. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/audit-config.prompt.mdManually: put the files in place
.gemini/commands/audit-config.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about audit-config
How do I install audit-config in Claude Code?
Add the marketplace with /plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor, then install the plugin: /plugin install ccam-config@claude-code-agent-monitor-plugins.
What does audit-config do?
It reads the Config Explorer overview and reports how many skills, agents, commands, MCP servers and hooks live in user and project scope, as a table with totals and a one-line verdict.
What does the command need to work?
A running Agent Monitor dashboard at http://localhost:4820. If nothing answers there, the command tells you to start it with npm start from the repo root.
Source text
audit-config.md · MIT
---
description: Quick Claude Code config audit — counts per surface (user vs project) and totals.
---
Run a fast configuration audit against the Agent Monitor Config Explorer at
`http://localhost:4820`. Fetch the overview in one call:
```bash
curl -s http://localhost:4820/api/cc-config/overview
```
The response is `{ roots, counts }` where `counts` includes:
`skills`, `agents`, `commands`, `outputStyles` (each `{ user, project }`),
`plugins`, `pluginsEnabled`, `pluginsDisabled`, `marketplaces`, `keybindings`,
`mcpServers` (`{ user, project }`), `hooks` (`{ user, project, "project-local" }`),
`memory`, and `settingsFiles`.
Print a compact one-screen table — one row per surface with User, Project, and
Total columns (sum the two scopes; for plugins/memory/etc. that have no scope
split, show the single total). Echo the resolved `roots.claudeHome` and
`roots.projectRoot` on the first line so the user knows what was inspected.
```
Config Audit (home=~/.claude project=/path/to/repo)
Skills .......... user 22 project 3 total 25
Agents .......... user 7 project 1 total 8
Commands ........ user 14 project 0 total 14
Output styles ... user 2 project 0 total 2
Plugins ......... 9 (enabled 7 / disabled 2)
Marketplaces .... 2
MCP servers ..... user 5 project 1 total 6
Hooks ........... user 7 project 0 project-local 0 total 7
Memory .......... 41 entries
Settings files .. 2
```
End with a one-line verdict: note the largest surface and any obvious sprawl
(e.g. heavy user-scope skills/commands). Cite only the numbers the API
returned — never fabricate counts. If the dashboard is unreachable, say so and
tell the user to start it with `npm start` from the repo root. Keep it to the
table + verdict; no extra prose.
© Son Nguyen · MIT · hoangsonww/Claude-Code-Agent-Monitor