Skills that let an AI assistant (Claude Code, Codex, Cursor, etc.) drive the Operon CLI on a user's behalf — install, mine sessions, polish, publish.
Add a skill to your local agent skills store via the npx skills CLI:
# Onboard a fresh user end-to-end
npx skills add Try-Operon/skills --skill operon-setup
# Help an existing user publish one of their drafts
npx skills add Try-Operon/skills --skill operon-publish
# Run a fresh mining pass on a recurring cadence
npx skills add Try-Operon/skills --skill operon-mineEach command drops a SKILL.md under ~/.claude/skills/<name>/
(and the equivalent path for Codex/Cursor).
| Skill | What it does |
|---|---|
operon-setup |
One-shot first-run: install the CLI, sign in via the browser, detect AI tools, pick a workspace, mine the first candidates. |
operon-publish |
Walks the user through picking one of their drafts and shipping it to their team. |
operon-mine |
Run a fresh mining pass against the user's Claude Code / Codex / Cursor sessions. Recurring loop. |
Operon also publishes these skills directly via the Cloudflare Agent-Skills Discovery RFC v0.2.0 at:
https://withoperon.com/.well-known/agent-skills/index.json
Each entry has a sha256 digest that matches the file in this repo.
Operon is a team layer for the AI workflows your engineers already
run. The CLI scans Claude Code, Codex, and Cursor sessions on each
laptop (local-first, redaction on disk), surfaces recurring prompts,
and publishes them to your team as installable operons — versioned
skills any teammate can operon install <workspace>/<slug>.
Read more at withoperon.com.
The source of truth for each skill lives in the Operon platform repo
under apps/landing/public/.well-known/agent-skills/. This repo is a
public mirror that the npx skills ecosystem can install from.
Open a PR here for typos / clarifications and we'll sync the change back to the source. For larger changes (a new skill, a behavior change), open a discussion at withoperon.com first.
MIT — see LICENSE.