skills.rest
Aki2022
Aki2022COMMUNITY
Explore all skills by this author →
GitHub

origin-skill-commonize

Unify multi-agent configuration files and skills into a single canonical source via symlinks.

Updated Dec 9, 2025
Search Tags:#claude code#multi-agent#agent configuration#codex cli#symlink#skill lint#canonical source
One-click install
npx skills add https://github.com/Aki2022/skills --skill origin-skill-commonize-aki2022
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: origin-skill-commonize
Source: https://github.com/Aki2022/skills/tree/main/origin-skill-commonize
Command: npx skills add https://github.com/Aki2022/skills --skill origin-skill-commonize-aki2022

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Multiple coding agents (Claude Code, Codex CLI, Gemini CLI, Antigravity) each read differently named config files (CLAUDE.md, AGENTS.md, GEMINI.md) and separate skill directories, causing duplicated content that drifts out of sync. This Skill establishes one canonical source under .agents/ and points every agent alias at it via symlinks, so any edit propagates to all agents immediately. ## Core Features & Use Cases - Canonical Symlink Topology: Consolidates global and per-repository agent configs, skills, commands, and non-secret MCP settings into .agents/ with per-tool subdirectories, keeping account-specific state (auth, sessions, history, cache) strictly separate. - Deterministic Skill Linting: Runs skill_lint.sh to validate SKILL.md presence, frontmatter name/description, directory-name matching, resource reference existence, broken symlinks, hook references, and pytest suites with exit-code reporting. - Topology and Mirror Auditing: Verifies global alias topology across Claude/Codex/Gemini/Antigravity accounts and checks third-party skill mirrors for byte-identical parity and freshness against mirrors.yaml. - Use Case: When setting up a new repository or noticing that Claude and Codex read different instruction files, use this Skill to inventory existing configs, merge divergent content, back up originals, and create symlinks so both agents share one canonical AGENTS.md. ## Quick Start Ask the agent to unify the CLAUDE.md and AGENTS.md files in this repository into a single canonical file using symlinks, then run the skill lint check to verify everything is consistent.

Frequently Asked Questions about origin-skill-commonize

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I share one AGENTS.md between Claude Code and Codex CLI?▼

Designate AGENTS.md (or .agents/skills for skills) as the canonical file, then create symlinks from each agent's expected path, such as CLAUDE.md pointing to AGENTS.md. The included unify_config.sh script handles backup, divergence detection, symlink creation, and verification.

How do I sync skills across Claude, Codex, and Gemini CLI?▼

Keep all skills in the canonical ~/.agents/skills/ directory and symlink each agent's skills path to it. Claude and Antigravity use a whole-directory symlink, while Codex and Gemini require one symlink per skill inside a regular directory.

Why is my canonical AGENTS.md not being read by Codex?▼

Codex reads AGENTS.override.md instead of AGENTS.md when both exist in the same directory, and the override replaces rather than extends the canonical file. Delete any AGENTS.override.md and verify reachability with a canary string test, since a correct symlink does not prove the file is read.

Can I symlink MCP settings or credentials between accounts?▼

Only MCP configurations verified to contain no secrets may be canonicalized under .agents/<tool>/, with credentials referenced by name or environment variable. Authentication tokens, sessions, history, and caches must never be symlinked and stay in per-account directories.

What does the skill lint check validate?▼

skill_lint.sh deterministically checks SKILL.md existence, frontmatter name and description, name-to-directory matching, existence of referenced scripts/references/assets paths, broken symlinks, hooks.json bash references, and runs pytest suites under scripts/tests. It reports results via exit codes without using an LLM.

What are the limitations of symlinking agent configs in git repositories?▼

Git stores symlinks as special blobs, but Windows environments with core.symlinks=false expand them into plain text files and break the links. Some CI systems and tools also do not follow symlinks, so critical paths need follow-up verification or an include-based alternative.

Related Skills

Explore More Software Engineering Skills→
richardmbaileyrichardmbailey

rb-setup-local-agent-skills

Audit local RB agent skill installations and symlink targets across Codex and Claude Code.

Community
Intermediate
bravegeekbravegeek

fate-character-generator

Generates complete Braille-ready Grade 2 translations from PDF, Word, or text input.

Community
Basic
likesjxlikesjx
1

ansible-skills-admin

Manage Ansible skill lifecycles across distributed gateways with drift cleanup.

Community
Intermediate
kaiiiiiiiiikaiiiiiiiii

nix-darwin

Configure macOS system defaults and services declaratively with Nix.

Community
Advanced