codex-fable5 — Claude Code skill: what it does, how to install
Applies Claude Fable 5 operating discipline to Codex agents. It enforces preliminary repo searches with rg, evidence tracking, and strict verification gates.
npx skills add baskduf/FableCodex --skill codex-fable5 -a claude-codeWhat codex-fable5 does
The skill imposes an explicit execution cycle on agent operations. Instead of editing files blind, the agent must inspect the workspace first using rg or rg --files. Routine single-step queries follow the standard flow, while complicated or review-heavy tasks activate gated decision maps.
Built-in Python scripts manage multi-stage operations using only the standard library. The codex_goals.py utility records milestone checkpoints, and codex_findings.py tracks unresolved review issues. If unresolved findings remain open, final verification stops immediately. Adding plugins/codex-fable5/bin to PATH enables terminal commands directly.
Prompt updates cannot alter underlying model weights, safety rules, or context boundaries. The skill enforces workflow structure rather than raw model abilities. For external Anthropic routing, make_litellm_config.py generates LiteLLM settings, but this still requires your own valid API key.
What it helps with
- translate Claude or Anthropic system instructions for Codex
- enforce repository inspection with
rgbefore any file modifications - track multi-step execution checkpoints using
codex_goals.py - log code review issues and block completion via
codex_findings.py
Example prompts for the agent
- Translate these Claude prompt guidelines into Codex format and generate an AGENTS.md file.
- Run a two-pass review of this directory and track any issues in
codex_findings.py. - Plan this refactoring using milestone gates in
codex_goals.py.
How to install codex-fable5
1Skills CLI (Vercel)
npx skills add baskduf/FableCodex --skill codex-fable5 -a claude-code2Manually: put the files in place
~/.claude/skills/codex-fable5/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/codex-fable5/
1Skills CLI (Vercel)
npx skills add baskduf/FableCodex --skill codex-fable5 -a codex2Manually: put the files in place
.agents/skills/codex-fable5/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add baskduf/FableCodex --skill codex-fable5 -a cursor2Manually: put the files in place
.cursor/skills/codex-fable5/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install baskduf/FableCodex codex-fable52Manually: put the files in place
.github/skills/codex-fable5/SKILL.md1Skills CLI (Vercel)
npx skills add baskduf/FableCodex --skill codex-fable5 -a gemini-cli2Platform CLI
gemini skills install https://github.com/baskduf/FableCodex --path plugins/codex-fable5/skills/codex-fable5Manually: put the files in place
.devin/skills/codex-fable5/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add baskduf/FableCodex --skill codex-fable5 -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about codex-fable5
How do I install codex-fable5 in Claude Code?
Run npx skills add baskduf/FableCodex --skill codex-fable5 -a claude-code. The skill file will be placed in ~/.claude/skills/codex-fable5/SKILL.md.
Does codex-fable5 modify model weights or context size?
No. It alters prompt-driven agent habits, tool selection, and verification steps. Core model weights, context limits, and safety filters stay completely untouched.
What helper scripts come with the skill?
It includes codex_goals.py for task tracking, codex_findings.py for review items, fable_coverage.py for parity checks, and make_litellm_config.py for generating LiteLLM configs.
