Set up a dialogue between this session and any other agent — create the dialogue file, emit one assume-nothing launch prompt for the other party, then play Party A yourself (write Turn 1 + self-drive). Run once, to start.
Interview me one question at a time to reach shared understanding of a feature before any plan is written. Each question comes with your recommended answer + the trade-off. No implementation plan until I say "alignment complete".
Dump the in-head topics of this session, triage each into a bucket (handover / memory / project-docs / do-now), act on them, and write a handover file for the next session to pick up. Say "/handover ok" to run non-interactively with default buckets.
View or update this project's session identity (name/color/model/effort) in .claude/session-identity.json — the file the cc launcher reads to name and color each session — then echo paste-ready /rename, /color, /model, /effort lines for the live session.
Disciplined root-cause analysis for diagnosing why something is broken. Use whenever the user is debugging a bug, regression, incident, outage, flaky or intermittent behavior, a failing test, or asks "why is X happening" / "why did this break" / "whats causing this" — even if they never say the words "root cause." It…