# Codex

> A Claude skill from knitli/toolshed.

- **Type:** Skill
- **Install:** `agentstack add skill-knitli-toolshed-codex`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [knitli](https://agentstack.voostack.com/s/knitli)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [knitli](https://github.com/knitli)
- **Source:** https://github.com/knitli/toolshed/tree/main/plugins/ctx/cross-client/.codex
- **Website:** https://knitli.com

## Install

```sh
agentstack add skill-knitli-toolshed-codex
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Context Hygiene Skill

## name
context-hygiene

## description
Audit and maintain AI context files across tool ecosystems. Detects stale claims, contradictions between CLAUDE.md/AGENTS.md/Serena/Cursor/etc., and validates references against the actual codebase.

## instructions

When activated, perform a context hygiene audit on the current repository:

### 1. Discovery
Scan for AI context files across all known tool ecosystems: CLAUDE.md, AGENTS.md, GEMINI.md, .claude/, .gemini/, .codex/, .cursor/, .serena/, .specify/, .roo/, .continue/, claudedocs/, specs/, plans/, docs/, info/, .mcp.json, .cursorrules, .windsurfrules, .continuerules, .clinerules, .aider.conf.yml

Report what you find as an inventory grouped by tool.

### 2. Staleness check
For each memory/instruction file, extract factual claims (paths, versions, symbols, counts, commands, dependencies) and validate them against the actual codebase. Report what's valid, what's stale, what's broken.

### 3. Drift detection
Compare memory files against each other to find contradictions and gaps. Flag where different tools have divergent views of the same project.

### 4. Summary
Produce a prioritized list of issues to fix, with specific file paths, line numbers, and recommended corrections.

### After code changes
When you make significant changes to file structure, dependencies, versions, or architecture, proactively check whether any context files need updating. The most important files to keep current are the root-level memory files (CLAUDE.md, AGENTS.md).

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [knitli](https://github.com/knitli)
- **Source:** [knitli/toolshed](https://github.com/knitli/toolshed)
- **License:** MIT
- **Homepage:** https://knitli.com

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-knitli-toolshed-codex
- Seller: https://agentstack.voostack.com/s/knitli
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.