Capafy logo Capafy
開始使用
PM Red Team & Discovery

PM Red Team & Discovery

by: Dave Dai
Stress-test any plan, PRD or decision: find the 3-5 assumptions that would kill it, with kill criteria and the cheapest test for each. Or turn a fuzzy idea into a discovery brief with opportunities, options, experiments and metrics.
#產品#顧問#生產力
評分
需要更多評價
已售出
0
使用方式
下載

What it does

PM Red Team & Discovery helps product managers, founders and strategy leads make better bets before money, people or a public commitment goes in. It works in two modes:

  • Red-Team: paste a plan, PRD, roadmap, pricing change, launch or decision. You get the 3-5 assumptions most likely to kill it, ranked by impact and likelihood, each with the strongest case for it, a kill criterion and the cheapest way to test it.
  • Discover: describe a fuzzy idea or problem. You get one clear outcome, 3-7 opportunities written as real user problems, 2-3 solution options each, the load-bearing assumptions, experiments with success thresholds, and an optional metrics brief (North Star, input metrics, guardrails).

Ask "should we build X?" and it runs Discover first, then red-teams the option it recommends.

Why it is different from a plain prompt

  • A scoring rubric (impact, likelihood, cheapness to test, 1-5 each) so the ranking is explicit, not vibes.
  • A 12-category assumption sweep (demand, value, competition, business model, channel, operations, technology, data, timeline, legal/policy, organisation, measurement) so blind spots get checked.
  • Steelman before attack: it argues the strongest version of each claim before challenging it.
  • No invented data: unknowns are labelled Unknown and turned into tests.
  • One question at a time: it only asks you about choices that change direction, and always proposes a recommended answer.
  • Ready-made templates and two complete worked examples are included.

Example

Input: "We run a meal-kit subscription. Plan: launch a $9/month add-on that lets members skip unlimited weeks without losing their discount. Goal: cut churn from 11% to 8% per month in one quarter."

Output (excerpt):

Rank Claim I x L (C) Kill criterion Cheapest test
1 Most churners leave because they need a break, not because of price 5x4 (5) "Need a break" < 30% of cancellations Query existing cancel-flow data (1 day)
2 Members will pay $9 instead of cancelling and re-joining 5x4 (4) < 3% of cancel-flow visitors click the offer Fake-door offer for 2 weeks

Followed by What Holds Up, What I Could Not Assess, and one Recommended Next Step.

Output format

Markdown in the language you write in: tables for the structured parts, short prose for the rest. If you ask for a file, it saves red-team-<topic>.md or discovery-brief-<topic>.md and also gives a short summary in chat.

How to use

Download the skill package and copy the pm-red-team-discovery folder into your agent client's skills directory (for example ~/.claude/skills/ in Claude Code; Codex, OpenClaw and other skill-compatible clients work the same way). No API keys, scripts or extra dependencies are needed.

Try:

  • "Red-team this launch plan: ..." (paste the plan or attach the PRD)
  • "Help me run discovery on this idea: ..."
  • "Should we build X? Here is what we know: ..."

FAQ

Does it need any API key or account? No. It runs entirely inside your own agent client.

Will it make the decision for me? No. It is decision support: it ranks risks and proposes tests and one next step. The call is yours.

Can it give legal or financial advice? No. It can flag that a legal or financial assumption exists and suggest how to verify it, but it does not replace professional advice.

What if I don't have data? That is the normal case. It labels the gaps and designs the cheapest experiment to fill them.


View more