daveangulo

twining-coordinate

Discovers agents, delegates tasks, and coordinates multi-agent work through capability matching. Use when orchestrating parallel work across multiple agents.

twining-decide

Records architectural and implementation decisions with rationale, rejected alternatives, and traceability. Use after making any non-trivial choice where alternatives exist.

twining-dispatch

Coordinates subagent dispatches through Twining — registers agents, posts delegations, and creates handoffs. Use when spawning subagents to keep all work visible in the coordination dashboard.

twining-handoff

Hands off work between agents or sessions via a committed handoff document plus a blackboard pointer. Use when ending a session with incomplete work, switching agents, or approaching context limits.

twining-map

Builds and queries the Twining knowledge graph — records entities, relations, and code structure. Use when documenting architecture or tracing dependencies between components.

twining-review

Reviews pre-commit state — verifies all significant decisions are recorded and traces decision coverage. Use before creating git commits to ensure Twining completeness.

LLM-judged semantic staleness review of Twining entries — flags decisions and blackboard entries tied to historical context that has moved on (dead sprints, renamed concepts, retired architectures). Opt-in, human-confirmed; complements deterministic staleness_review. Use when housekeeping is clean but old entries still reference concepts that no longer make sense.

twining-verify

Runs pre-completion verification — checks unresolved warnings, decision hygiene, and drift. Recommended for complex tasks; for simple tasks a status post via twining_post is sufficient.