bmad-product-brief

Invoke the skill with create intent and it guides you through a discovery conversation, capturing your idea, stakes, and source material before drafting. It writes brief.md with YAML frontmatter into a run folder and logs every decision to a memlog for auditability.

Use update mode: the skill reads the existing brief, addendum, and memlog, then reconciles the change signal against prior decisions before editing. It surfaces conflicts with earlier decisions and logs overrides rather than silently patching.

Yes, validate mode critiques the brief against its own stated purpose, citing specific lines and caveating what cannot be evaluated. Findings return inline and the skill always offers to roll them into an update.

Yes, headless mode completes the detected intent without prompting, inferring from provided materials and the existing workspace. If intent stays ambiguous it halts with a blocked JSON status and a reason field instead of asking questions.

It produces brief.md as the main artifact, an optional addendum.md for depth that belongs downstream, and a .memlog.md append-only audit trail. All files live in a run folder under the configured brief output path.

Skip it when you only need a quick one-page summary with no discovery or audit trail, since the coaching process invests time in surfacing assumptions. Deep market sizing or exhaustive competitive teardowns are also out of scope and routed to a dedicated research skill.