Skip to content

Command Clusters

Commands are organised into clusters (subcommand counts shown). This is an overview — the full catalog lists all 190.

Cluster Subcmds Purpose
agents 9 Agent-layer lifecycle — init/optimize/audit AGENTS.md, persona interview
fix 8 Targeted fixers — ci, pr-comments, quality, refs, portability, seeder, route
roadmap 6 Roadmap authoring + autonomous execution (step/phase/full), materialize
council 6 Poll external AIs for a neutral second opinion — default, pr, design, …
optimize 6 Optimization sweeps — skills, agents-dir, rtk filters, project, prompt
memory 6 Engineering-memory lifecycle — add/load/propose/promote/mine
analyze 6 Retrospective/prospective — postmortem, premortem, decision, incident
brand 5 Brand-as-UX — strategy, identity, tokens, voice, review
feature 5 Feature workflow — explore, plan, refactor, roadmap, dev
ghostwriter 5 Public-figure voice profiles (mandatory disclosure)
video 5 AI video pipeline — from-script, from-song, scene, storyboard, stitch
team 4 Governed cross-model access — review, adversarial, delegate, status
tests 4 Test authoring/execution — create, execute, e2e-plan, e2e-heal
worktree 4 Governed git worktrees — create, status, verify, cleanup
knowledge 4 Local knowledge ingestion — ingest, list, forget, cross-repo
judge · tdd · image · profile · sync 3 each LLM-judge loops · TDD cycle · character images · session profiles · config sync
analytics · bug · challenge-me · context · cost · module · override · package · post-as · pr · project · research · review · team-knowledge 2 each Focused two-command clusters

Plus ~22 standalone commands (no cluster): work, implement-ticket, jira-ticket, refine-ticket, estimate-ticket, agent-handoff, agent-status, mode, orchestrate, condense, humanize, threat-model, prepare-for-review, upstream-contribute, and more.