Skills

Skills (recipes) are slash commands that define what to do. Each skill is described in a SKILL.md file with frontmatter metadata and instructions for agents. Brewcode includes 8 skills covering the full task lifecycle — from project analysis to E2E testing.

On top of those there is a dynamic ninth skill: it is not shipped with the plugin, it is generated into every target project by /brewcode:superreview and invoked there as /superreview, tailored to that project’s tech stack, agents, rules and tracker. See the emitted skill page for details.

Summary table

#	Skill	Command	Model	Purpose
1	agents	/brewcode:agents	opus	Creates, improves, syncs Claude Code subagents
2	convention	/brewcode:convention	opus	Extracts etalon classes, patterns, architecture into convention docs
3	e2e	/brewcode:e2e	opus	Orchestrates e2e testing: BDD scenarios, Playwright autotests
4	rules	/brewcode:rules	sonnet	Syncs KNOWLEDGE.jsonl or session learnings to project rules
5	skills	/brewcode:skills	opus	Lists, improves, creates, syncs Claude Code skills
6	spec	/brewcode:spec	opus	Creates SPEC.md task spec via research + interaction
7	superreview	/brewcode:superreview	opus	Generates a project-tailored deep-review skill: domain-expert routing + scope discipline + mechanical gates + adversarial validation
8	teams	/brewcode:teams	opus	Creates and manages dynamic teams of domain agents
★	superreview (emitted)	/superreview	opus	Per-project deep review — written into your repo by /brewcode:superreview

Delegation — one contract, all eight skills

Every skill here fans work out to subagents, and all of them follow the same sizing rule.

One subagent = ONE bounded unit: one deliverable, roughly 5 files, roughly 10 steps. Anything bigger is split into N tasks and spawned in ONE message, in parallel. That is why convention runs 10 layer agents instead of one, spec partitions research into 5-10 areas, and teams creates agents one file per spawn.

Every spawn prompt carries six fields — a bare one-line task is never enough:

Field	Content
GOAL	the overall task and why it exists — the point beyond the file edit
ROLE	what this agent owns; what it must NOT touch
SCOPE	exact paths/commands in bounds + explicit out-of-bounds
CONTEXT	what is already done, by whom, what runs in parallel — trimmed to what THIS agent needs
CONSUMER	who or what uses the result next, and the shape it must fit
DONE	acceptance criteria + the exact report shape you want back

The same brief lives in the runtime templates these skills read — the e2e mode references, convention’s rules guide, spec’s SPEC-creation reference, and superreview’s emitted skill template — so a fan-out three levels deep still gets a full brief.

Lifecycle

The recommended skill execution order for a standard project cycle:

Skill dependencies

spec --> superreview --> /superreview (emitted into your repo)
  |
  +-- convention (optional)
  +-- rules
  +-- teams (optional)

Independent Skills

These skills work standalone — no prerequisite steps required: