brainstorming
Interactive clarification before any feature, spec, or design work — when the request lacks persona, success signal, or scope. Asks one grounded question at a time, informed by `browzer explore`/`search` on the actual repo, optionally dispatches parallel research agents (WebFetch, WebSearch, Firecrawl, Context7) for unknowns, hands off to generate-prd. Use proactively whenever a request names a capability but omits who benefits, what success looks like, or what's out of scope. Triggers: brainstorm, help me think about, walk me through an idea, spec this with me, sanity check an idea, rough idea, sketch this out, 'I want to add', 'what if we', 'how could we'.
Repository SourceNeeds Review
code-review
Post-implementation team review of a feature's diff. Spawns 4 mandatory agents in parallel — senior-engineer (cyclomatic complexity, DRY, clean code, best practices), software-architect (system design, race conditions, clean architecture, caching, performance), qa (regressions, edge cases, butterfly-effect breakage), regression-tester (runs scoped tests over modified files + their browzer deps) — plus domain specialists discovered via /find-skills. Every agent gets the diff + browzer deps (forward + reverse) + browzer mentions and may run browzer explore to detect prior art / duplication. Read-only — `receiving-code-review` applies fixes next. Triggers: code review, review this feature, audit my changes, review the diff, post-implementation review, team review, peer review, find issues in this PR.
Repository SourceNeeds Review
commit
Write a Conventional Commits v1.0.0 message mirroring the repo's last 5 commits, stamp the `on-behalf-of: @browzeremb` org-attribution trailer (per GitHub's organization-commit doc — orgs link via `on-behalf-of:`, NOT `Co-authored-by:` which is for human collaborators), and run `git commit`. Reports the SHA. Does NOT push. Use whenever the user wants to commit staged changes. Triggers: commit, commit this, save this, checkpoint, finish this task, ship this commit, write a commit message, conventional commit.
Repository SourceNeeds Review
execute-task
Implement one task end-to-end by dispatching domain specialists per its `task.explorer.skillsFound[]`. Each specialist loads project skills first, writes code scoped to `task.scope`, reports gates + invariants, and aggregates into `task.execution`. For free-form requests without a plan, calls `generate-task` first. Tests are NOT authored at this phase — they're written after `code-review` + `receiving-code-review` by the `write-tests` skill. Triggers: execute TASK_03, run the first task, implement task 02, do this task, ship TASK_N, build the feature from the plan, 'implement this'.
Repository SourceNeeds Review