spwf

Cross-cutting — Audit both the project .claude/ and personal ~/.claude/ for quality issues. Covers CLAUDE.md scope and length, agent descriptions, skill frontmatter, and settings.json conflicts. Always audits both scopes. Produces a prioritised P1/P2/P3 fix list.

Phase 2 — Approve plan. Reviews the task list from spec for quality (atomicity, testability, clarity), then applies four adversarial lenses (Skeptic, Architect, Minimalist, Security) as advisory input for the human reviewer. Quality issues are blocking; adversarial findings are advisory. Security lens identifies tasks touching auth, billing, user input, or secrets so the builder has explicit awareness before writing those tasks. Presents everything for explicit human go/no-go before building starts.

Phase 3 orchestrator — Full Red-Green-Refactor cycle. For each pending task: invokes write-tests (Red: write failing tests), then opsx:apply (Green: implement via OpenSpec), then run-tests (Verify: confirm green before moving on). Loops until all tasks complete. After all tasks complete, runs opsx:verify (spec sign-off), then recommends simplify (Refactor).

Pre-phase orchestrator — accepts any input (issue tracker ticket, Slack message, file, or freeform description), classifies it as a bug or a change, then routes to the appropriate path. Bug path runs systematic root-cause investigation and produces todo/BUG-{slug}.md. Change path runs a lightweight qualification check and produces todo/{slug}.md. Both outputs feed /spwf:challenge.

Gate — Challenge a plan, ideation file, or design relentlessly. Accepts a file path as $ARGUMENTS (defaults to the most recent file in todo/ if omitted). Reads the file first, interviews until all open questions are resolved, then runs a scope-sizing check — recommends splitting into multiple changes if the work spans independent boundaries, or proceeding as one change if tightly coupled.

7 hooks across 2 events