From forwward-teams
Runs lint, typecheck, build, and test checks; iteratively reads errors and fixes them up to 4 times to verify code works before commits or deploys.
How this skill is triggered — by the user, by Claude, or both
Slash command
/forwward-teams:gateThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Run checks, read errors, fix them, repeat. Prove code WORKS by executing it.
Run checks, read errors, fix them, repeat. Prove code WORKS by executing it.
Read .claude/project.json for build commands. If missing, init from the plugin CLI:
${CLAUDE_PLUGIN_ROOT}/scripts/cli init
Or run checks directly via the plugin CLI:
${CLAUDE_PLUGIN_ROOT}/scripts/cli gate # lint + typecheck + build + test
${CLAUDE_PLUGIN_ROOT}/scripts/cli gatekeep -g # same, with PASS/FAIL report
${CLAUDE_PLUGIN_ROOT}/scripts/cli gatekeep -l # lint only
${CLAUDE_PLUGIN_ROOT}/scripts/cli gatekeep -t # test only
Run up to 4 iterations:
| Error Type | Fix | Don't |
|---|---|---|
| Type errors | Fix the type, add the import | Use @ts-ignore |
| Build errors | Fix imports, exports, modules | Skip the check |
| Lint errors | Fix the actual issue | Blanket disable |
| Test failures | Fix the code or the test | Delete the test |
NEVER change business logic during gate. Only fix types, imports, lint.
After 4 iterations without full pass:
| Flag | What |
|---|---|
-l | Lint only |
-c | Typecheck only |
-b | Build only |
-t | Tests only |
-a | App startup (dev server + health check) |
-u | UI tests (Playwright/Cypress) |
-g | Full gate: lint + typecheck + build + test |
--all | Everything including app + UI |
npx claudepluginhub iankiku/forwward-teamsRuns quality gates for linting, type checking, unit tests with coverage, spec compliance, and smoke checks at local/CI/deploy levels.
Runs phased verification checks for build, types, lint, tests, security scans, and git diffs in JS/TS/Python projects. Invoke after changes or before PRs.
Enforces test gates, evidence-based verification, and frontend safety checks before deployment. Blocks deploys that fail tests or lack evidence.