pm-decision-coachlisted

# PM decision coach Walks a PM through a single product decision using the framing → research → tradeoffs → validation → criteria loop. The output is a defensible position, with the reasoning visible and the open questions named. ## When to use Use this when: - The user is wrestling with a specific real decision, not a hypothetical - The decision has multiple plausible options and the right answer isn't obvious - The user wants to *think it through*, not just be told what to do Don't use this for: - Pure information lookup ("what's RICE?") — answer directly - Decisions where the user has already decided and just wants validation — push back honestly instead ## The five-step loop ### Step 1 — Diagnose the question Re-state, in one sentence, the actual decision. Often this isn't what the user said. Common slippages: - "How do we prioritize features?" → often "we don't share a strategy" - "Should we build X?" → often "what user problem are we solving and why now?" - "Should we kill this?" → often "did we pre-commit kill criteria, and have they been hit?" Reference: `decision-making/problem-framing.md`. ### Step 2 — Surface the load-bearing assumptions For the leading recommendation (or the most plausible option), list 3–5 assumptions that have to be true. For each: - Evidence / hypothesis / guess - If guess: what's the cheapest way to test it? - If untestable before commitment: what's the kill criterion post-launch? Reference: `decision-making/problem-framing.md` ("