diff --git a/README.md b/README.md index 1706166..4a24a60 100644 --- a/README.md +++ b/README.md @@ -140,16 +140,18 @@ No registry commands, no publishing steps, no terminal knowledge beyond paste. T Each team member creates skills from their own domain and shares them. Over months the organization accumulates a library of reusable skills: - Sales team shares `/sales-report-skill` -- Engineering shares `/deploy-checklist` -- Legal shares `/quarterly-compliance` -- Data science shares `/customer-churn-model` -- SRE shares `/incident-runbook` +- Engineering shares `/deploy-checklist-skill` +- Legal shares `/quarterly-compliance-skill` +- Data science shares `/customer-churn-skill` +- SRE shares `/incident-runbook-skill` Any colleague installs any skill with one `git clone`. Any agent on any platform can invoke it. Knowledge compounds instead of evaporating. -### For power users: the registry CLI +### For teams and consultants: the skill registry -If your organization prefers a centralized catalog (browsable, searchable, versioned), the registry tooling is available: +When your organization has more than a few skills, the agent offers to set up a **team skill registry** — a single repo where all skills are published, browsable, and installable. Think of it as an internal app store for agent skills. + +The agent sets it up automatically when it detects you're building for a team. Or you can set it up manually: ```bash python3 scripts/skill_registry.py init --name "Acme Corp Skills" @@ -159,7 +161,9 @@ python3 scripts/skill_registry.py search "sales" python3 scripts/skill_registry.py install sales-report-skill ``` -This is optional. Most teams start by sharing `git clone` links and add the registry when the skill count grows. +The registry is a git repo. Clone it once, and every team member can browse and install any skill. No servers, no databases — just git. + +**For AI consultants:** Install agent-skill-creator, create skills from your client's workflows, set up the registry, and hand over a self-sustaining system. After you leave, the team keeps creating and sharing skills on their own. --- diff --git a/SKILL.md b/SKILL.md index 1a37e50..3d73e82 100644 --- a/SKILL.md +++ b/SKILL.md @@ -290,6 +290,80 @@ The goal: the user who created the skill sends a one-liner to their colleague on **If the user says no**, that's fine — the skill is already installed locally and working. They can always share later. +### Set Up a Team Skill Registry (Consultant / Org Rollout) + +When a user is setting up skills for a team or organization — not just sharing a single skill — offer to create a **centralized skill registry**. This is the infrastructure that lets the organization accumulate and discover skills over time. + +``` +It looks like you're building skills for a team. Want me to set up a +shared skill registry? This gives your organization a single place to +publish, browse, and install all skills — like an internal app store. +``` + +**If the user says yes, do all of this automatically:** + +1. **Initialize the registry**: + ```bash + mkdir -p /path/to/team-skills-registry + python3 scripts/skill_registry.py init --registry /path/to/team-skills-registry --name "Company Name Skills" + ``` + + Use a sensible path — suggest the user's preferred location, or default to `~/team-skills-registry`. + +2. **Create a remote repo for the registry** (same GitHub/GitLab detection as skill sharing): + ```bash + cd /path/to/team-skills-registry + git init && git add -A && git commit -m "feat: Initialize team skill registry" + + # GitHub + gh repo create team-skills-registry --private --source=. --push + gh repo edit --add-topic agent-skill-registry + + # Or GitLab + glab repo create team-skills-registry --private --defaultBranch main + git remote add origin && git push -u origin main + ``` + + Note: the registry repo should be **private** by default (it's an internal catalog). + +3. **Publish the just-created skill to the registry**: + ```bash + python3 scripts/skill_registry.py publish ./sales-report-skill/ --registry /path/to/team-skills-registry --tags sales,reports + ``` + +4. **Show the team how to use it**: + ``` + Registry set up! Here's what your team needs to know: + + Clone the registry (one time): + git clone ~/team-skills-registry + + Browse available skills: + python3 scripts/skill_registry.py list --registry ~/team-skills-registry + + Search for skills: + python3 scripts/skill_registry.py search "sales" --registry ~/team-skills-registry + + Install a skill (auto-detects platform — VS Code Copilot, Cursor, etc.): + python3 scripts/skill_registry.py install sales-report-skill --registry ~/team-skills-registry + + Publish a new skill: + python3 scripts/skill_registry.py publish ./my-new-skill/ --registry ~/team-skills-registry --tags tag1,tag2 + ``` + +**When to offer registry setup:** +- User mentions "team", "organization", "department", "colleagues", "company" +- User creates multiple skills in one session +- User asks about sharing or distributing skills at scale +- User is clearly an AI consultant or admin setting up infrastructure + +**When NOT to offer registry setup:** +- User is creating a single personal skill +- User already shared via the one-liner git clone flow and is satisfied +- A registry already exists (check for `registry/registry.json` or `~/team-skills-registry`) + +The registry is a git repo. It gets version history, access control (repo permissions), and review workflows (PRs) for free. No servers, no databases, no new tools to learn. + See `references/pipeline-phases.md` for detailed Phase 5 instructions. ### Generated SKILL.md Format