Durable repo memory and optional automated queue orchestration. Use when the user explicitly says "koder-pattern", "with koder-pattern", or "use koder-pattern"; asks to install/bootstrap it; explicitly requests a durable koder/ artifact; or asks to build/run/recover a koder queue or Harnex workflow. Do not trigger for ordinary coding, planning, review, research, or repo opening unless durable koder output or orchestration is requested.
End-of-session hand-off for this repo; commit reviewed intentional work, refresh koder/STATE.md, and do not report success until Git is clean.
Start-of-session hand-off for this repo; inspect durable state and live Git facts, then render a compact Past/Present/Future report.
Simulated Council of High Intelligence for structured multi-persona deliberation. Use when the user explicitly asks for /council, Council of High Intelligence, a council/triad/duo debate, or named council personas for decision analysis.
Audible TTS notifications for agent workflow status. Use when the user asks for speak, voice, audio, or TTS notifications, or when a repo workflow explicitly requests voice notifications during long runs, harnex dispatches, queue runs, test results, commits, open/close handoffs, or surprising blockers. Do not use for routine replies or sensitive content.
UI/UX design intelligence adapted for BFBB/Holm/Zippy apps and raw browser UIs. Use when designing, building, reviewing, fixing, or polishing app pages/components, responsive layout, accessibility, motion, color, typography, charts, or design systems.
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.
Create, rewrite, and improve agent skills with strict structure, explicit trigger wording, and repeatable evaluation. Use when the user asks to build a new skill, merge existing skills, refactor a SKILL.md, or improve when a skill should and should not trigger.