Scans a project's current practice, classifies each convention by how it is best enforced, and routes it to the fittest mechanism - config, project doc, rule, a project skill, or a paste-ready pointer. Use when setting up an existing project so the first agent execution follows its conventions, or when asked to establish, capture, or align project conventions.
Runs a post-task retrospective - detects user corrections, failures, and repeated mistakes, stages lessons as candidates, and proposes consent-gated improvements routed to their fittest destination (project rule, entry-file fact, memory, backlog, or skill update). Use when a task is complete or about to be marked done, before committing or closing out, or when the user says done, wrap up, or ship it.
Audits any skills directory on three layers - format against shared authoring conventions, content facts (commands, APIs, versions) against live documentation, and rules that never changed the model's behavior - reporting per-layer findings with sources. Use when asked to audit, review, or check whether skills are outdated, stale, bloated, or still correct.
Sets up and re-syncs a project's entry-file harness block — the wiring that brings retro in — plus cross-agent interop glue, with idempotent, diff-first, marker-managed writes; retro owns the learning loop. Use when setting up AI collaboration for a new or existing project, or when re-syncing the managed block after the skills were updated. Optional — projects bootstrap without it.
Creates, edits, or removes a single project rule in the agent's native rule format, enforcing a non-discoverability admission filter and a hard rule budget with provenance stamping. Use when the user asks to add, change, or delete a project rule, or when another skill hands over a drafted rule.
Implement tasks from an OpenSpec change. Use when the user wants to start implementing, continue implementation, or work through tasks.
Archive a completed change in the experimental workflow. Use when the user wants to finalize and archive a change after implementation is complete.
Enter explore mode - a thinking partner for exploring ideas, investigating problems, and clarifying requirements. Use when the user wants to think through something before or during a change.