Thanks for your interest in contributing. This marketplace ships to six agentic harnesses (Claude Code, OpenAI Codex CLI, Cursor, OpenCode, Gemini CLI, GitHub Copilot) from a single Markdown source.
- AGENTS.md — canonical context (table of contents)
- ARCHITECTURE.md — top-level architectural map
- docs/authoring.md — portable-content style guide (read this before adding new components)
- docs/harnesses.md — per-harness capability matrix
- docs/plugin-eval.md — quality evaluation framework
- Create
plugins/<name>/with.claude-plugin/plugin.json. - Add agents in
agents/, commands incommands/, skills inskills/. - Update
.claude-plugin/marketplace.jsonwith your entry. - Naming: lowercase, hyphen-separated. Never use
__(the adapter namespace separator). - Run
make generate-allto refresh the committed native-install registries (CI gates registry drift). - Run
make validateandmake gardento surface any issues before submitting.
Full frontmatter conventions in docs/authoring.md.
Every PR runs these on CI (.github/workflows/); run them locally before pushing:
make validate STRICT=1 # structural validation across all harness outputs
make garden STRICT=1 # drift, dead-link, stale-artifact detection
make test # full pytest suite (plugin-eval + tools/tests/)
make smoke-test # real-CLI subprocess tests (OpenCode, Gemini, Codex, Claude)Code-quality checks (also in CI):
cd plugins/plugin-eval
uv run ruff check ../../tools/ src/plugin_eval/
uv run ruff format --check ../../tools/ src/plugin_eval/
uv run ty check ../../tools/ src/plugin_eval/Your content ships to five harnesses — some have stricter conventions than Claude Code:
- Codex hard-truncates skill bodies at 8 KB. Keep
SKILL.mdshort; push detail intoreferences/details.md. - OpenCode requires lowercase tool names. Don't write
`Read`inline — write "open the file" or use the lowercase form. - Cursor doesn't honor per-agent
tools:allowlists — use it as a hint only. - Copilot maps Claude model aliases (
opus/sonnet/haiku) to the GPT-5 family; agentdescriptionmust be a plain string. - All harnesses use ≤150-line context files. Don't bloat
AGENTS.md/CLAUDE.md.
plugin-eval's harness_portability dimension catches most of these mechanically;
read docs/authoring.md for the full guide.
- Open an issue first (template-driven). Use the appropriate issue template.
- Fork the repo, branch from
main. - Make changes; run quality gates.
- Open a PR referencing the issue.
- CI must pass; reviewers approve; squash merge.
- Bugs / features / new components: use the GitHub issue templates.
- Code of Conduct violations: see
.github/CODE_OF_CONDUCT.md. - Discussions: https://github.com/wshobson/agents/discussions.