orqit

Use when: implementing one scoped issue, building/deploying/validating a product, capturing QA evidence, and giving a Ship/Revise recommendation.

Use when: orchestrating PLANNER → BUILDER → REVIEWER across one or more issues to reduce founder context-switching. Spawns role agents as subagents and routes their outputs through named gates.

Use when: triaging ideas, prioritizing work, writing specs, reviewing content quality, and producing issue breakdowns for implementation.

Use when: evaluating visual/audio quality, defining style guides, reviewing business health, auditing pipeline state, or running /status.

Structured 5-whys postmortem template for severity-gated mistakes, with canonical output marker and required destinations.

Living registry of platform operations that are known-automatable (API/CLI with session-store receipts) vs. truly manual (with evidence of why). Consult before labeling any step "manual" or "founder web-UI" in a plan, spec, or handoff.

Unified session lifecycle — start, resume, monitor, and end Copilot CLI sessions. Covers context management, twin sync, issue state, and branch cleanup.

Degraded-but-disciplined mode for when Claude Opus 4.7 premium requests are exhausted or must be conserved. Codifies the recipe that recovers ~85-92% of Opus quality on standard-tier models via extra process discipline.

Weekly refresh trigger for the Model Routing Matrix. Compares current Copilot CLI version to the Matrix's "Last verified" pin and produces a draft refresh issue when they diverge. Founder reviews the draft and merges matrix updates; this skill never auto-creates issues or edits the matrix.