/rota-decide
Record a hard-boundary decision with explicit forbids and permits.
Manual only: you type it, other skills never start it.
When to use
Use on “decide on X”, “we’re committing to X”, “lock in the boundary that Y”, or when a session has produced a constraint future work must respect.
Reach for another skill when
- you cannot state forbids and permits, so it is a gotcha rather than a commitment:
/rota-learn - the question is still open and needs an experiment first:
/rota-spike
Draft for topic "Architecture": Rule: never store session tokens client-side Why: <from the conversation> Forbids / Permits: <drafted from context> Write it? (nothing is written without your "Write it")
- a four-part entry (rule, why, forbids, permits) in .rota/DECISIONS.md
Flags and modes
--from-learning <topic>- seed rule and why from a KNOWLEDGE.md bullet
--from-spike <name>- seed rule and why from a spike finding (inconclusive spikes are refused)
--supersede <topic> "<title>"- replace an existing decision with a new one
--retire <topic> "<title>"- mark a decision lifted, nothing replacing it
Details
Captures a hard-boundary decision into .rota/DECISIONS.md. Manually confirmed, never auto-invoked. Unlike learnings in KNOWLEDGE.md, decisions are commitments with explicit forbids/permits; /rota-work, /rota-debug, /rota-refactor, /rota-plan, /rota-brainstorm, /rota-vision and /rota-review consult them as constraints. --from-learning <topic> promotes a hardened KNOWLEDGE.md bullet into a decision (rule/why are pre-filled; you supply the forbids/permits), and --from-spike <name> promotes a .rota/spikes/<name>.md finding the same way (inconclusive spikes are refused). See decisions for the full flow.