fullreviewlisted

# /fullreview — Comprehensive Multi-Domain Review Dispatch parallel agent teams across every domain that matters, consolidate findings, fix CRITICAL/HIGH issues, and loop until clean or the iteration budget is exhausted. ## Usage ``` /fullreview # full pipeline on current project (default: 5 loops) /fullreview --loops 3 # cap iterations at 3 instead of 5 /fullreview --max-loops 5 # same thing /fullreview --scope security # single-domain review, still looped /fullreview --scope backend,jobs # multi-domain, only listed /fullreview --no-fix # report only, never edit code /fullreview --fast # skip the 5 "extra" domains (perf, observability, etc.) ``` ## Required behavior The agent running this skill must: 1. **Invoke** `superpowers:dispatching-parallel-agents` and `superpowers:verification-before-completion` skills up-front (they govern parallel dispatch and the "claim clean only with evidence" gate). 2. **Track progress** via TaskCreate/TaskUpdate with one task per loop. Mark in_progress when starting, completed when done. Never claim "clean" without running the verification command in the same message. 3. **Capture baseline** before any review: detect and run the project's typecheck and test commands (e.g. `npm run typecheck && npm test`, `pytest`, `cargo test`, `go test ./...`), save exit codes, save test counts. Abort with a clear message if the baseline is already broken and tell the user to fix the baseli