By burnjohn
Two execution agents usable with or without a spec-driven workflow: `implementer` (build one task and self-verify) and `test-writer` (add or extend tests without changing production behavior).
Use proactively to implement ONE task/slice from an implementation plan (or one clearly scoped request). Discovers the project's stack and conventions before writing anything, applies whichever skills fit the task type, and self-verifies with the project's existing tests + typecheck before finishing. Safe to run in parallel as long as each instance owns non-overlapping paths.
Use proactively to add or extend tests for existing code — backend services, domain/business logic, or UI components and hooks. Discovers the project's test stack and split conventions first, writes only test files, and self-verifies by running the suite + typecheck with pasted evidence before finishing.
AI Demo Marketplace — a public Claude Code plugin marketplace built for the AI Agentic Engineering workshop. It's a catalog you add once, then install shared plugins from: skills, agents, hooks, and MCP servers.
The catalog lives separately from any product — its own owner, its own release cadence. A product release must not force a plugin release, and vice versa.
This first commit is the baseline. It ships the marketplace structure and docs with an empty plugin catalog. Real plugins, a validation script, CI, and a catalog website are added stage by stage during the workshop — each on its own branch.
/plugin marketplace add burnjohn/ai-demo-marketplace
/plugin install <plugin-name>@ai-demo-marketplace
<plugin-name> is the name of an entry in .claude-plugin/marketplace.json; ai-demo-marketplace is the catalog name. The baseline ships with no plugins yet — the catalog fills up as the workshop progresses.
| Command | What it does |
|---|---|
/plugin marketplace update | Refreshes the catalog — pulls the current list of available plugins/versions |
/plugin update <plugin>@ai-demo-marketplace | Updates an already installed plugin to a new version |
Refreshing the catalog does not update installed plugins by itself.
.claude-plugin/marketplace.json # the catalog (plugin registry) — empty in the baseline
plugins/ # each plugin in its own subdirectory (added during the workshop)
docs/PLUGIN-GUIDELINES.md # plugin anatomy (manifest, dependencies)
docs/SECURITY.md # permissions, secrets policy
docs/RELEASES.md # SemVer, tags, update, rollback
CODEOWNERS · CONTRIBUTING.md
The baseline is deliberately minimal so we can build it up live. Later branches add, roughly in order:
scripts/validate-marketplace.mjs) — secrets / absolute-path / escape checks..github/workflows/validate.yml) running the linter + claude plugin validate.Before adding a plugin, read CONTRIBUTING.md and docs/PLUGIN-GUIDELINES.md.
SKILL.md compatible (Markdown + YAML frontmatter) with other agents.Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub burnjohn/ai-demo-marketplace --plugin code-agentsTesting skill pack. Currently React Testing Library conventions for component and hook tests with Vitest — project-agnostic, no framework assumptions beyond React.
Frontend engineering skill pack for React 19 + Next.js 15 (App Router): code architecture/placement, React best practices, and Next.js patterns. TypeScript now lives in `typescript-skills`; React Testing Library in `testing-skills`.
Spec-Driven Development workflow: spec -> plan -> implement -> test -> verify -> retro. Ships the authoring and verification agents, the multi-agent retrospective skill, and the slash commands that drive the chain. Execution agents come from `code-agents`.
Language-level TypeScript skill pack: type-level programming, strictness settings, utility types, and a diagnostic script. Stack-agnostic — useful in frontend, Node backend, and library projects alike.
Read-only research agent that locates, gathers, and fact-checks information from the codebase or the public web and returns it in a structured, cited format. Never edits files.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Evidence-gated AI coding workflow: scan → analyze → plan → TDD → execute → fix → verify → review, powered by Codebase Memory MCP >= 0.9.0 with optional Serena LSP intelligence. Includes blast-radius planning, test/cycle gates, independent review, and Windows Git Bash hook auto-resolution.
Harness-native ECC plugin for engineering teams - 67 agents, 279 skills, 94 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Browser automation and end-to-end testing MCP server by Microsoft. Enables Claude to interact with web pages, take screenshots, fill forms, click elements, and perform automated browser testing workflows.
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.