What exactly does the Codex Home Audit skill check?

It audits your Codex home directory for control-plane drift, risky state, and cleanup opportunities across configuration, agents, hooks, skills, plugins, and telemetry settings.

How does it handle sensitive information like API keys?

The skill includes safety constraints to redact secrets and sensitive data, such as tokens, API keys, and .env contents, by default.

What happens if the audit finds a critical error?

The tool will write the report but exit with a non-zero status, highlighting the blocker issues like missing configuration files or invalid global guidance sources.

Will this skill automatically modify my Codex configuration?

No, the skill is report-first by design. It generates a diagnostic report and only applies changes if you explicitly request an implementation pass.

Can I use this to audit a non-default Codex directory?

Yes, you can provide an optional path for 'codex_home' to audit any directory instead of the default ~/.codex location.

Codex Home Audit

Name: Codex Home Audit
Author: jscraik

byjscraik

•

보안 및 테스팅

Audits Codex home directories to detect configuration drift, security risks, and cleanup opportunities.

The Codex Home Audit skill provides a comprehensive health check for your Codex environment, focusing on the integrity of the control plane. It identifies critical issues such as instruction precedence conflicts, stale shadow files, and telemetry risks across configurations, agents, hooks, and plugins. By generating a detailed, dated Markdown report, it allows developers to visualize their system state and plan cleanups without the risk of automated changes. It is an essential tool for maintaining high-performance, secure agentic workflows and ensuring that global rules and local profiles remain synchronized and effective.

주요 기능

01Audits telemetry and retention postures for data privacy compliance

02Reviews automation surfaces including hooks, agent roles, and local state

03Detects instruction precedence issues and guidance drift across profiles

04Identifies configuration risk hotspots and deprecated instruction files

052 GitHub stars

06Generates actionable Markdown reports with specific next-step recommendations

사용 사례

01Cleaning up stale history, logs, and invalid plugin manifests in the home directory

02Resolving conflicting AI instructions across global and local agent profiles

03Hardening security by auditing shell wrapper rules and sensitive data exposure

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jscraik/agent-skills codex-home-audit

For use in Claude.ai and ChatGPT

주요 기능

01Audits telemetry and retention postures for data privacy compliance

02Reviews automation surfaces including hooks, agent roles, and local state

03Detects instruction precedence issues and guidance drift across profiles

04Identifies configuration risk hotspots and deprecated instruction files

052 GitHub stars

06Generates actionable Markdown reports with specific next-step recommendations

사용 사례

01Cleaning up stale history, logs, and invalid plugin manifests in the home directory

02Resolving conflicting AI instructions across global and local agent profiles

03Hardening security by auditing shell wrapper rules and sensitive data exposure

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jscraik/agent-skills codex-home-audit

For use in Claude.ai and ChatGPT