Development

Use when encountering any bug, test failure, build failure, or unexpected behavior, before proposing any fix. Covers root-cause investigation, hypothesis testing, and the escalation rule when repeated fixes fail.

Use to review any corpus document at any time, and always before a gate on the four artifacts no doc_standards covers — the architecture spine, SRS, SDD, and SPEC. Reads the lens set from the component's risk_accepted, dispatches bmad-review, and stamps the `review-trace` trace on those four only. Not for code review.

Use when someone needs numbers about this project — progress for a client update, an estimate before the work is committed, task rows ready to paste into a tracker, or a self-contained brief/PRD deliverable. Four intents, progress and estimate and dispatch and render. Never invents a number.

Use before a gate, or after a batch of changes, to find drift between .what, .how, and .control — against each other and against the rules in .constitution. Scoped to what the gates already passed have actually produced, and to what each component's mode actually demands. Read-only — it reports, it never edits.

Use when something cannot be decided now and must not be silently assumed. Files it into one of four lists in .control/questions/ by what the reader has to do about it, and closes it in place when the answer arrives.

Clean up completed closed specs by archiving to `.archive/specs/` or pruning from disk, wrapping lifecycle.py with fail-closed safety. Invoke as `/wdi-prune-or-archive [--spec <id>|--all-closed] [--archive|--prune] [--dry-run]` or bare `/wdi-prune-or-archive` for interactive selection.

Use at G2 Product — when a PRD is created or an existing promise changes. Two intents, prd and update. Checks position, dispatches bmad-prd, verifies the result against prd-guide.md, and lands the memlog. Never writes the PRD itself.

Use at G1 Problem — when the product brief is created, updated, or validated. Checks position and preconditions, dispatches bmad-product-brief, then verifies the result against brief-guide.md and the template. Never writes the brief itself.

Use when a fact from outside the code has to be recorded — a meeting that finished, or a non-technical fact that constrains what may be built. Two intents, meeting and fact. Routes decisions and open questions to their own skills.

Use for anything that must exist before work can start or continue — scaffolding the registries at install, birthing Product Components after G2, setting or changing a component's mode, setting or reviewing its risk_accepted, refreshing the two structure maps, repairing the engines after an install or update, and writing this product's inventory readers. Seven intents. Never writes .what/ or .how/ content beyond a skeleton.

Check project delivery status, open or pending specs and tickets, current gate progress, and determine what to build or which skill to invoke next. Answers from this project's status registry and five gates.

Use when the owner has to decide something — an open question, a defect, a design fork, a failing validator, a vague worry — and wants the reading done for them. Investigates, then briefs in six fixed sections, in the owner's language. Writes no file; the decision itself goes to wdi-decision or wdi-question.

Use when a decision worth remembering has been made, when one has to be carried into the documents it governs, or when a planning assumption turns out to be void. Three intents — open, accept, apply. Wraps bmad-correct-course. Produces a globally numbered DEC-.

Sync to development branch, prune merged worktrees/branches, configure target application smoke environment, and build a test checklist from closed tickets. Invoke as `/wdi-daily-what-to-test [web <target>|mobile <target>|desktop]`.

Turn raw manual-test notes into a triaged, reviewed spec/ticket ready for wdi-autopilot to pick up in a separate session. Invoke as `/wdi-daily-what-to-build [reviewer] [--no-review] <your raw notes>`.

Compose and launch the autonomous daily loop routine (default 10m interval) with self code-review and peer-review runners resolved from local configuration. Invoke as `/wdi-daily-autopilot [self-review] [peer] [interval] [--skip-peer-review|--no-review]` (`[self-review]` alias `[in-session]`).

Use at G4 Component — the depth of one Product Component, as deep as that component's mode and no deeper. Two intents, behaviour and design. Owns .what/<pc>/ slots 02-05 and .how/<pc>/ minus 01-ux. Skipped entirely at mode catalog.

Use at G5 Release — one spec from open to closed in one supervised run. Opens the spec, hands the owner to-spec and to-tickets, ships every ticket to a green PR through a five-step pipeline, then closes the spec. One invocation, not four.

Use at G3 Blueprint — the one whole-product portrait, written once. Two intents, catalog and platform. Owns the use case catalogue, actors, domain model, cross-component business rules, the glossary, the spine, C4, cross-cutting, and the three inventories. Never writes a single component's depth.

Use when the owner wants the agent to carry the product from where it stands to every FR delivered, without being asked a question in between. Two doors — a preflight that ends in one mandate the owner accepts, then unattended iterations fired by a loop, each working as far as it safely can. One run, one branch, one PR. Every decision the agent takes lands in one ledger the owner reviews in parallel, never as a prompt.

Review pull requests (PRs), branches, commit ranges, and related PR stacks for consequential defects. Use for requests such as "review PR 42" or "review this branch", including PR URLs and owner/repo#number references. Produces evidence-backed findings and coverage gaps; does not implement fixes.

面对未知软件、程序、服务、二进制、协议或系统时，系统建立理解并持续修正模型的研究方法。当需要侦察一个陌生目标的结构、行为、数据流、网络交互、认证机制或验证关于它的假设时使用。通用方法论，不绑定任何特定软件、语言或框架。

Use when the user wants to hand off, transfer, pause, or continue the current session in a new session or with another agent — asks for a "session handoff", a "prompt para a próxima sessão", to "continuar de onde paramos", or invokes /session-handoff; also when context is running low and in-flight work must survive a session boundary. Produces one self-contained, copy-pasteable prompt.

Use when branch work is finished and the user asks to close the session out — "/session-end", "encerra a sessão", "fecha essa branch", "vamos encerrar aqui", "abre o PR e faz o merge", "leva até o merge e limpa a branch", "finaliza e limpa o worktree". Also OFFER it (do not auto-run it) when the user signals closure without naming it — "deixa tudo pronto", "o que falta a gente fazer?", "zera esses pendings". NOT for work still in progress, NOT when the user only asked to push, and NOT for bui...