Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
134 lines (93 loc) · 6.16 KB

CONTRIBUTING.md

File metadata and controls

134 lines (93 loc) · 6.16 KB
Raw

Contributing to TeamAI CLI

Thanks for your interest in improving TeamAI! This document explains how to get a dev environment running, how to structure changes, and how to get your PR merged.

Development Setup

Fork Tencent/teamai-cli, then clone your fork:

git clone https://github.com/<your-user>/teamai-cli.git
cd teamai-cli
git remote add upstream https://github.com/Tencent/teamai-cli.git
npm install

Common commands

npm run build          # Build with tsup → dist/
npx tsc --noEmit       # Type check
npm run lint           # oxlint; CI fails on any warning
npx vitest run         # Run unit tests
npx vitest run --coverage
npm run test:e2e       # Build and run E2E tests; remote cases require credentials

npm run lint needs Node ^20.19 or >=22.12 (oxlint's requirement); the CLI itself still supports Node 20.

Running your local build

npm run build && npm link
teamai --version

Dogfood against a public team repo

All contributors should exercise pull, hooks, and recall while developing the CLI. Create a regular (not template) public team repo under teamai-hub, e.g. https://github.com/teamai-hub/teamai-cli-dev, with main branch protection (PR + review). Put only public-safe skills / rules / docs there.

In the CLI clone (not as teamai init .):

npm run build && npm link
teamai init https://github.com/teamai-hub/teamai-cli-dev --scope workspace --role dev
teamai pull
git status   # nothing under .teamai/ or tool dirs should be staged for this repo

Pitfalls:

  • Init the canonical hub URL, not a personal fork. teamai init <url> treats that URL as the team repo; a fork diverges immediately, and GitHub push/PR today targets the configured remote (no fork-to-upstream flow).
  • Do not run teamai init .. That is single-repo mode: it turns the CLI source tree into the team repo and writes scaffolding at the repo root (easy to commit by mistake).
  • Push failed (you can push manually later) on member registration is expected without write access. Local config is still saved; teamai pull still works.

Write access and team stats

digest / dashboard read stats/, sessions/, and members/ from the team repo. Those files are written via git, so no write ⇒ not in team stats.

Giving every internet contributor write on the hub repo is not acceptable.

Who Hub repo access Required setup In team digest
Contributors read init + pull no
Collaborators (after a few PRs) write, main protected full, including reports yes

Project Layout

src/
  providers/         # git hosting provider abstraction
    github/          # GitHub (gh CLI or GITHUB_TOKEN)
    gitlab/          # GitLab (GITLAB_TOKEN)
    gitcode/         # GitCode (GITCODE_TOKEN)
    cnb/             # CNB (cnb CLI or CNB_TOKEN)
    tgit/            # Tencent TGit (gf CLI)
    git/             # any Git host, no platform API
  resources/         # per-resource-type handlers (skills, rules, docs, env, ...)
  utils/             # shared helpers (git, fs, logger, prompt, ...)
  *.ts               # top-level command entry points (init, push, pull, ...)

See docs/dev/adding-a-provider.md for how to add a new git provider.

Before you change src/session-flow/, read docs/dev/agent-sessions.md. It lists the rules that span more than one file.

Before you change src/workflow/, read docs/dev/workflow-engine.md. It lists the rules that span more than one file.

Making a Change

  1. Update from upstream/main, then create a feature branch. Prefer a git worktree for code changes when practical.
  2. Write tests for your change (we target 80%+ coverage).
  3. Run npx vitest run, npx tsc --noEmit and npm run lint — all must pass.
  4. Use conventional commits where possible: feat:, fix:, chore:, docs:, refactor:, test:.
  5. Push the branch to your fork and open a PR against Tencent/teamai-cli:main. Describe the problem, the fix, and anything reviewers should pay attention to.

Your PR also gets an informational Code Erosion report (SlopCodeBench verbosity/erosion metrics) posted as a comment — it never blocks the merge and is just there to flag creeping complexity. See docs/dev/ci-code-erosion.md.

Coding Style

  • TypeScript strict mode is on; avoid any unless genuinely needed.
  • Prefer async/await over callbacks.
  • Keep commands in src/*.ts thin — heavy lifting lives in src/resources/ or src/utils/.
  • Avoid narrating comments ("// increment counter"). Comments should explain why, not what.

Testing Guidelines

  • Unit tests go in src/__tests__/. Mirror the source file name (init.ts → init.test.ts).
  • Mock external I/O (git, fetch, child_process) at the module boundary.
  • Avoid relying on real network access unless guarded by an env variable (like TEAMAI_TEST_TOKEN).

Iterate on the affected E2E files; run the full suite when changing the E2E runner, shared fixtures, or test isolation:

npm run test:e2e -- <test-file> [-t "<test-name>"]   # add --maxWorkers=2 on constrained machines

Workers share the prepared dist/ and OpenCode binary: never rebuild or reinstall them inside a test. Cases needing remote credentials skip without them (CI E2E setup). PRs changing runtime behavior (not docs- or tests-only) include one real-CLI verification, such as a focused E2E run, with its command and result.

Bug Reports & Feature Requests

Please file issues at github.com/Tencent/teamai-cli/issues. Include:

  • What you tried to do
  • What happened (error output, stack trace)
  • What you expected
  • Your OS, Node.js version, and teamai --version

Security

For security issues, please do not open a public issue. Email the maintainers or use GitHub's private vulnerability reporting.

License

By contributing, you agree your contribution will be licensed under the MIT License.