Multi-lens code review of a diff: six specialized lenses (spec-conformance, correctness, security, performance, design, tests) merged into one deduplicated, severity-ranked verdict. Fast single-pass (min) and orchestrated ensemble (max) modes.
1 skill · 5 commands · 7 agents
Record a before/after test drive of a UI or backend API change: /test-drive:baseline records the app before you edit, /test-drive:compare replays the same scripted scenario after the change and gives a deterministic pass/warn/fail verdict, with a side-by-side video for UI or a request, response, and server-log report for API.
2 skills
Prove a finished change with replayable, claim-first evidence: 1 to 5 falsifiable claims, one artifact per claim (Playwright video and trace for UI, a Playwright request trace or a vhs tape or a Hurl transcript for an API, a script run for a CLI or library), and a summary of PROVEN, FAILED, or NOT-EVIDENCED. Provides the show-evidence skill.
1 skill
Turn a natural-language description of a flowchart, architecture diagram, process map, or state machine into an editable tldraw document (.tldr) plus a rendered PNG/SVG. Provides the /tldraw:diagram skill.
1 skill
Claude Code mod: a band line showing who is doing the work this turn (main chat, helper agents, models, tasks, status, skills), plus a /agent-watch pane with per-agent blocks, per-model totals and a timeline.
Claude Code mod: a band line showing how full the chat is, saving a HANDOFF.md at a threshold (auto or on request), then clear and continue or resume in a new session, plus a /auto-handoff command.