Skip to main content
skill CX Codex

config-explorer — Claude Code skill: what it does, how to install

Inspects and edits Claude Code and Codex settings through CCAM with timestamped backups. Useful for auditing hooks, memory, rules, and MCP servers.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add hoangsonww/Claude-Code-Agent-Monitor --skill config-explorer -a claude-code
Open on GitHub

What config-explorer does

This skill inspects and modifies configuration surfaces in Claude Code and Codex using CCAM. It handles checks for MCP servers, agent memory, hooks, user skills, and instruction files. Reading Claude Code state relies on CLI calls like ccam config claude overview or ccam config claude skills --scope user.

Editing follows a strict server allowlist. Claude Code writes accept only text artifacts and keybindings, requiring a JSON payload alongside the --yes flag. In Codex, config.toml remains edit-only, while new profiles use ccam config codex profile <name> --yes.

Every write and deletion generates an automatic timestamped backup. The skill prohibits overwriting redacted previews and blocks editing state files outside the allowlist. Without the ccam CLI installed, none of these commands will work.

What it helps with

  • inspect active MCP servers and hooks in Claude Code
  • audit agent memory and user-scoped skills
  • update Codex config.toml without risking state files
  • create new configuration profiles with automatic backups

Example prompts for the agent

  • Run ccam config claude overview to inspect the Claude Code configuration.
  • Check registered hooks and MCP servers with ccam config claude mcp.
  • Create a new Codex profile using ccam config codex profile dev --yes.

How to install config-explorer

1Skills CLI (Vercel)

npx skills add hoangsonww/Claude-Code-Agent-Monitor --skill config-explorer -a claude-code

2Manually: put the files in place

~/.claude/skills/config-explorer/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/config-explorer/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about config-explorer

How do I install config-explorer in Claude Code?

Run npx skills add hoangsonww/Claude-Code-Agent-Monitor --skill config-explorer -a claude-code in your terminal. You can also inspect it at ~/.claude/skills/config-explorer/SKILL.md.

What does config-explorer do?

It provides inspection and allowlisted editing of Claude Code and Codex configurations through CCAM. Each write or delete action produces a timestamped backup before applying changes.

What safety limits apply when modifying files?

Claude Code writes are restricted to text artifacts and keybindings via JSON files with --yes. Redacted previews cannot be overwritten, and state files outside the allowlist remain locked.

Source text

SKILL.md · MIT
---
name: config-explorer
description: >
  Inspect and safely edit the Claude Code and Codex configuration surfaces
  exposed by CCAM. Use when auditing skills, agents, commands, plugins,
  marketplaces, MCP servers, hooks, settings, memory, keybindings, profiles,
  rules, or instruction files, and when a backup-backed allowlisted edit is
  required.
---

# Config Explorer

## Claude Code

Read with:

```bash
ccam config claude overview
ccam config claude skills --scope user
ccam config claude mcp
ccam config claude hooks
ccam config claude memory
```

Supported writes are limited to text artifacts and keybindings. Use a JSON file
and `--yes`. Every write or delete creates a timestamped backup.

## Codex

Read with:

```bash
ccam config codex overview
ccam config codex read /absolute/path
```

Use `edit` only for the server's unredacted editable allowlist. Supported writes
and deletes are backup-backed. Base `config.toml` is edit-only. Profiles are
created with `ccam config codex profile <name> --yes`.

Never overwrite a redacted preview. Read the editable source before saving.
Never edit plugin or MCP state files outside the server's allowlist.

© hoangsonww · MIT · hoangsonww/Claude-Code-Agent-Monitor

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set60

daily-standup Turns your recent Claude sessions into a standup report: work done per project, errors hit, what it cost. CX Codex +1 ★ 1,121repo stars mcp-server Configures, builds, and runs the CCAM MCP server across Claude Code and other hosts. Useful when choosing transport modes or locking down mutation policies. CX Codex +1 ★ 1,121repo stars hook-setup Checks and installs CCAM monitoring hooks for Claude Code and Codex. Use it when adding a provider, repairing hooks, or verifying an installation. CX Codex +1 ★ 1,121repo stars history-portability Pulls in Claude or Claude history and shifts whole CCAM datasets between machines. Useful when you switch laptops or need a restore. CX Codex +1 ★ 1,121repo stars webhook-management Configure CCAM webhook targets, connect them to alert rules, and inspect delivery records. Use the skill when managing notification endpoints from the CLI. CX Codex +1 ★ 1,121repo stars remote-collection A skill that walks the agent through CCAM remote data sources: connect machines over SSH and pull Claude and Codex history. Useful when sessions live beyond your laptop. CX Codex +1 ★ 1,121repo stars alert-management Inspects fired CCAM alerts and manages monitoring rules for token thresholds, inactivity, and status duration. CX Codex +1 ★ 1,121repo stars session-compare Compare two Agent Monitor sessions across cost, tokens, events, and agent activity. Use it to see what changed between runs with different outcomes. CX Codex +1 ★ 1,121repo stars

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor