Sync Skills Across Multiple AI Agent Installations
Distributes an authored skill to all 4 local agent skill folders - Codex, Claude Code, Pi, and Hermes - keeping them in sync.
Why it matters
Keep agent skills synchronized across four different AI agent platforms on a local MacBook by managing a canonical skill repository with symlinks and selective copying, ensuring every agent has access to the same capabilities without manual duplication.
Outcomes
What it gets done
Author skills once in the canonical ~/.agents/skills/ folder and verify symlinks
Copy updated skills to Hermes agent while Claude and Pi auto-sync via symlinks
Validate all four agent locations show identical skill file byte counts
Detect and warn about broken symlinks or orphaned skills in wrong directories
Install
Add it to your toolbox
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/ag-distribute-skill-to-all-agents | bash Overview
Distribute a Skill Across All Agents
Distributes an authored skill across all four local agent skill folders - Codex, Claude Code and Pi via symlink, Hermes via manual copy - so every agent can discover it. Use it to sync a skill globally across agents, not for project-specific skills (which belong in the repo) or single-agent edits (patch that file directly instead).
What it does
Distributes a locally-authored skill across all 4 agent skill locations on a MacBook so every agent can discover it: Codex/OpenAI Agents at ~/.agents/skills/ (canonical - author here first), Claude Code at ~/.claude/skills/ (a symlink to ~/.agents/skills/, so it's auto-covered), Pi Agent at ~/.pi/agent/skills/ (also a symlink to ~/.agents/skills/, auto-covered - the path is nested under /agent/, not ~/.pi/skills/), and Hermes Agent at ~/.hermes/skills/ (an independent copy, the only location needing a manual copy).
When to use - and when NOT to
Use it when a skill should be made available across multiple local agent skill folders, or when the user asks to sync or distribute skill updates to other agents. Do NOT use it for a project-specific skill - that belongs in ./.claude/skills/, ./.pi/agent/skills/, etc. inside the repo, not globally, since a project-local skill overrides the global one on collision. Also don't use it to edit just one agent's skill (e.g. a Hermes-only workflow) - patch that file directly instead of propagating it everywhere. Removing a skill globally is destructive: first show the exact skill directories that would be removed, confirm with the user, then use their preferred safe deletion method for ~/.agents/skills/ and ~/.hermes/skills/.
Inputs and outputs
Author the skill in ~/.agents/skills/<skill-name>/SKILL.md first (the canonical location). Verify the .claude symlink is intact with a one-time check (ls -la ~/.claude/skills should show ~/.claude/skills -> ~/.agents/skills; if it's a real directory instead, the copies have diverged - ask before touching). Then copy only to .hermes, since .claude and .pi are symlinks already covered:
SKILL=<skill-name>
cp -r ~/.agents/skills/$SKILL ~/.hermes/skills/
Finally verify all 4 locations show identical byte counts for SKILL.md; if .claude or .pi shows a different count, that symlink is broken and needs investigating before proceeding. Updating an existing distributed skill uses the same flow - re-copy to .hermes/skills/ (the .claude and .pi symlinks update automatically) - but cp -r overwrites by default, so use rsync -a --delete instead if the skill folder has nested files that may have been removed:
rsync -a --delete ~/.agents/skills/$SKILL/ ~/.hermes/skills/$SKILL/
Integrations
Covers four skill-loading agents and their known pitfalls: ~/.pi/skills/ is the wrong location for Pi (it loads only from ~/.pi/agent/skills/; anything found in the wrong path is an orphan, confirm with the user before deleting); ~/.claude/skills is a symlink, not a folder, so an explicit cp -r into it errors with "are identical" and should be skipped; project-local skill folders (./.pi/agent/skills/ or .pi/skills/ inside a repo) override the global one on collision and this skill only handles global distribution; .pi/agent/skills is a symlink to .agents/skills and shouldn't be cp'd into directly, while .hermes/skills remains the one independent copy that shouldn't be unilaterally consolidated into a symlink unless the user asks; a newly-distributed skill won't appear inside an already-running Hermes session until restart, since Hermes snapshots skills at session start, though it works immediately for future sessions and the other 3 agents; and SKILL.md filename casing matters on case-sensitive volumes, so it must stay uppercase.
Who it's for
Users running multiple local coding agents (Codex, Claude Code, Pi, Hermes) side by side who want one authored skill kept in sync across all of them, rather than maintaining four separate copies by hand.
Source README
Distribute a Skill Across All Agents
When to Use
- Use when a skill should be made available across multiple local agent skill folders.
- Use when the user asks to sync or distribute skill updates to other agents.
The user has 4 agent skill locations on his MacBook. A skill must exist in each (or via symlink) to be discoverable by every agent.
The 4 Canonical Locations
| Agent | Skills Folder | Notes |
|---|---|---|
| Codex / OpenAI Agents | ~/.agents/skills/ |
Canonical - author skills here first |
| Claude Code | ~/.claude/skills/ |
Symlink → ~/.agents/skills/ - writing to .agents/skills automatically covers Claude |
| Pi Agent | ~/.pi/agent/skills/ |
Symlink → ~/.agents/skills/ - auto-covered. (Path is /agent/ nested - NOT ~/.pi/skills/) |
| Hermes Agent | ~/.hermes/skills/ |
Independent copy - the only one needing a manual copy |
Workflow
- Author the skill in
~/.agents/skills/<skill-name>/SKILL.md(canonical). Followeffective-agent-skillsSKILL.md guidance. - Verify the
.claudesymlink is intact (one-time check):
If it's a real directory instead of a symlink, the user has diverged copies - ask before touching.ls -la ~/.claude/skills # Expect: ~/.claude/skills -> ~/.agents/skills - Copy to
.hermesonly (.claudeand.piare symlinks - already covered):SKILL=<skill-name> cp -r ~/.agents/skills/$SKILL ~/.hermes/skills/ - Verify all 4 locations show identical byte counts:
All four numbers must match. Iffor p in ~/.agents/skills/$SKILL ~/.claude/skills/$SKILL ~/.pi/agent/skills/$SKILL ~/.hermes/skills/$SKILL; do echo "$p: $(wc -c < $p/SKILL.md) bytes" done.claudeor.pishows a different byte count, that symlink is broken - investigate before proceeding.
Updating an Existing Distributed Skill
Same flow - re-copy from ~/.agents/skills/ to .hermes/skills/. The .claude and .pi symlinks update automatically. cp -r overwrites by default; use rsync -a --delete if the skill folder has nested files that may have been removed:
rsync -a --delete ~/.agents/skills/$SKILL/ ~/.hermes/skills/$SKILL/
Pitfalls
~/.pi/skills/is the wrong location. Pi Agent loads from~/.pi/agent/skills/only. A skill placed in~/.pi/skills/is invisible. If you find skills already there, they're orphans - confirm with the user before deleting.~/.claude/skillsis a symlink, not a folder.cp -r ~/.agents/skills/foo ~/.claude/skills/will error with "are identical". Skip the explicit Claude copy.- Project-local skills exist too -
./.pi/agent/skills/(or.pi/skills/) inside a repo overrides the global one on collision (later-discovered wins). This skill only handles GLOBAL distribution. .pi/agent/skillsis a symlink →.agents/skills. Don'tcpinto it (errors "are identical"); it auto-syncs. Only.hermes/skillsis an independent copy - don't unilaterally consolidate Hermes into a symlink unless the user asks.- Hermes snapshots skills at session start. A newly-distributed skill won't appear inside a running Hermes session until restart (it works fine for future sessions and for the other 3 agents immediately).
- Filename casing matters on case-sensitive volumes.
SKILL.mdmust be uppercase.
When NOT to Use This Skill
- Skill is project-specific → put it in
./.claude/skills/,./.pi/agent/skills/, etc. inside the repo, not globally. - Editing one agent's skill only (e.g. a Hermes-only workflow) → patch that file directly, don't propagate.
- Removing a skill globally is destructive. First show the exact skill directories
that would be removed, confirm with the user, then use the user's preferred
safe deletion method for~/.agents/skills/and~/.hermes/skills/.
Limitations
- Adapted from
davidondrej/skills; verify local paths, tools, credentials, and agent features before acting. - For commands, remote access, scheduling, browser automation, or file-changing workflows, get explicit user approval and confirm the target environment first.
FAQ
Common questions
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.