Doc-driven implementation engineer. Reads SDLC design docs first, verifies all library APIs via Context7, then writes clean production code. Enforces anti-slop rules — no over-engineering, no defensive bloat, no hallucinated APIs. Use after design docs exist.
Gauntlet loop — multi-agent quality harness: a lead sets a real reference bar (a named product, test suite, or baseline to match or beat), splits the goal into gradeable units, builders produce artifacts in clean context, and blind fresh-per-round critics…
Image-quality gating and single-annotation tooling for captured screenshots (img-gate.mjs, annotate.mjs). Placeholder skill front door — T21.2 scope only.
Claude-Design-style visual iteration loop — render the running UI, screenshot at 375/768/1440, critique against docs/design/tokens.json and design principles, apply fixes, re-capture until clean (cap 3 iterations). Also extracts a token baseline from an…
Ralph Wiggum deep-onboard: Step D4 — re-run focused HANDOFFs for rows flagged by /onboard-verify. Does NOT re-run the full DISCOVER phase.
OWASP audit, threat modeling, CVE/dependency scanning. Supports --quick (default, ~10 min) and --deep (Ralph Wiggum: exhaustive OWASP x SAST rules x iterative attack chain, ~45-90 min). SAST runs on Opengrep + in-house bpm-rulepacks (never license-restricted…
Reflow the module-ticket graph so a newcomer can pick up work. Reads plan.json + completed artifacts, marks done modules, recomputes the claimable (ready + unclaimed) set, flags write-scope collisions, and emits a HANDOFF for a claimed module. Read-mostly and…
Project intelligence steward — audits CLAUDE.md / AGENTS.md alignment with actual codebase, captures session learnings, updates project docs. Use after major sessions or when docs feel stale.