Agent Workshop

Ready-to-use AI agents and skills for Claude Code and Codex, extracted from real projects and packaged as installable plugins. You get review agents that catch problems in specs, tests, and code, plus workflow skills for handoffs and reports — without building any of it yourself.

Install

This repo doubles as a plugin marketplace — a catalog Claude Code can install plugins from directly. In a Claude Code session, add it once:

/plugin marketplace add giostriquer/agent-workshop

Then install one (or both) of the plugins:

/plugin install toolkit@agent-workshop
/plugin install agent-workshop@agent-workshop

Using Codex instead:

codex plugin marketplace add giostriquer/agent-workshop
codex plugin add toolkit@agent-workshop
codex plugin add agent-workshop@agent-workshop

For Cursor, the plugins ship in the Cursor plugin format (.cursor-plugin/). Import this repo as a Team Marketplace (Teams/Enterprise, admin): Dashboard → Settings → Plugins → Team Marketplaces → Add Marketplace → Import from Repo, point it at giostriquer/agent-workshop, then install toolkit / agent-workshop from Customize in the sidebar. (Plugins published to the official cursor.com/marketplace install with /add-plugin or the Add to Cursor button.)

The plugins

toolkit — use right away

Review agents and direct-use skills, ready immediately after install with nothing to configure:

• Agents: spec-reviewer (design specs and plans), code-quality-reviewer (maintainability and structure of a diff), test-quality-reviewer (test code), pattern-reviewer (code-pattern conformance), vigil (your agent/skill setup itself), and ci-watcher (your branch's PR CI). All read-only — they inspect and report, never edit your files.
• Skills: handoff-review, handoff-pr, and handoff-goal hand in-flight work to a fresh session or agent; doc-to-html turns a markdown report into a polished dark-themed HTML page; claim-check runs an unbiased investigation of a premise (ticket, hunch, or question) and returns a verdict plus a readiness dossier; qa-sweep fans a QA team over a broad surface and corroborates every finding firsthand before it counts; code-quality-review runs an unusually strict, structure-first maintainability review over a branch's diff and pushes for restructurings that delete complexity; get-pr-comments triages the active PR's review comments into a prioritized action list (read-only).

Details in plugins/toolkit/README.md.

agent-workshop — adopt the full scaffold

One guided skill, agent-workshop-onboard, for adopting the project-coupled scaffolding into your own repo — the agents that need adapting to your project (profile slots, conventions) and the workflow skills meant to live in-repo, plus the conventions that tie them together. It inspects the target, produces a read-only adoption plan, and only copies files after you approve. The direct-use review agents and self-contained skills don't need this — install toolkit for those.

Going deeper

• docs/agents/ and docs/skills/ — the origin story of every agent and skill: what problem it solved and how it's used in practice.
• docs/conventions/ — the portable working rules the agents rely on.
• docs/adoption/README.md — the pack catalog, maturity labels, and host support.

Everything here came from months of lived-in use on real projects, sanitized to be portable. It's not a framework or a methodology — adopt the pieces that earn their keep in your project and skip the rest.