thinkroom-trigger

Thinkroom trigger

Use this skill to decide whether a question deserves a Thinkroom research job.

Invoke Thinkroom when

All of these are true:

1. The outcome matters. A shallow answer could lead to a costly, hard-to-reverse, or strategically important decision.
2. The answer is genuinely uncertain. There are competing hypotheses, interpretations, designs, or risk models.
3. Independent perspectives add value. The question benefits from branches that form views without seeing sibling outputs.
4. Traceability matters. The user needs evidence, provenance, verification status, dissent, or an auditable synthesis.

Typical examples include architecture choices, product strategy, investment theses, incident hypotheses, policy trade-offs, due diligence, and research synthesis.

Do not invoke Thinkroom when

• the task is a trivial lookup, deterministic calculation, simple rewrite, or direct file/tool operation;
• one authoritative source or one bounded verification can settle the question;
• the user needs execution rather than research;
• additional perspectives would add latency but no meaningful discrimination.

Handoff

When the trigger passes:

1. Load thinkroom-operate.
2. Frame one explicit question plus only the context every branch should share.
3. Choose the smallest branch count that covers distinct hypotheses; the default is three.
4. Submit with an idempotency key when retries are possible.
5. Treat evidence verification status as authoritative and preserve unresolved dissent.

Thinkroom supports decisions; it does not replace live verification, domain authority, or approval for external effects.