Index of every page on this site: llms.txt
All skills
am-will avatar

/create-hook

@9f954c3
by am.willam-will/codex-skills1k stars
60
Run it yourself

Create or update Codex hooks for global or project scope by asking for scope and hook type when ambiguous, then writing the hook config, scripts, and any required feature flag config.

  • 3 files
  • 3 KB
  • Updated 8 months ago
  • GitHub
Use this Skill: https://skilld.dev/gh/am-will/codex-skills/create-hook

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈49 tokens for metadata: the name and description. ≈463 when used: this file. ≈254 more on demand in 2 files.

Description uses 2.5% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈197 characters, or 2.5%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Create Hook

Use when the user wants to add Codex hooks and explains the behavior they want.

Workflow

  1. Determine scope.

    • If the user says global, use the personal/home scope.
    • If the user says project, use the repo scope.
    • If scope is missing or unclear, ask: Should this be global or project-scoped?
  2. Determine hook type.

    • If the event is missing or unclear, ask which hook they want:
      • SessionStart
      • UserPromptSubmit
      • PreToolUse
      • PostToolUse
      • Stop
    • If the user wants more than one, keep them separate unless they explicitly want shared behavior.
  3. Determine behavior.

    • Ask what the hook should do if the intent is not already clear.
    • Prefer the smallest deterministic script that satisfies the request.
  4. Write the hook files in the chosen scope.

    • Global: ~/.codex/hooks.json and ~/.codex/hooks/
    • Project: <repo>/.codex/hooks.json and <repo>/.codex/hooks/
    • Keep script paths explicit and relative where practical.
  5. Enable the feature flag if the target Codex build still gates hooks.

    • Set [features] codex_hooks = true in the relevant config.toml.
  6. Keep the hook output clear.

    • Use statusMessage for a short visible signal.
    • For logging hooks, write to a repo-local or home-local log file.
    • For blocking hooks, return a nonzero exit only when the hook should actually stop the tool.
  7. Verify the result.

    • Run a command that should trigger the hook.
    • Confirm the log, output, or blocking behavior matches the requested scope and event.

Practical defaults

  • Use PreToolUse for prevention.
  • Use PostToolUse for logging, formatting, or test follow-up.
  • Use UserPromptSubmit for prompt shaping and injected context.
  • Use SessionStart for startup context.
  • Use Stop for end-of-turn checks.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No third-party reports yet.

Signed by skilld at 9f954c3. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last month.

Steadyupdated 8 months ago

README badge

README badge for am-will/codex-skills/create-hook