assumption-mapping
List the assumptions behind an idea and rank them by importance and evidence
Install / Use
npx skills add deanpeters/Product-Manager-Skills --skill assumption-mappingInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Customer SupportSupported Platforms
Our assessment of assumption-mapping
assumption-mapping scores 86/100 on our quality scale, 208th of 319 Customer Support skills we index.
Its SKILL.md is 12 KB long, well organised into 31 sections and no code examples: a thorough specification that gives an agent plenty to work with.
With 7,063 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated about 3 months ago, so assumption-mapping is actively maintained.
- No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
- Its trust signals score 88/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-11-16. Automated pattern scan on 2026-11-16. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
assumption-mapping compared with similar skills
All 4 of these similar skills score higher than assumption-mapping; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| assumption-mapping (this skill)by deanpeters | 86 | 7.1k | 3mo ago | SKILL.md |
| x402by browser-use | 100 | 121.0k | 3d ago | SKILL.md |
| algorithmic-artby anthropics | 100 | 177.9k | 55d ago | SKILL.md |
| pptxby anthropics | 100 | 177.9k | 55d ago | SKILL.md |
| designby nextlevelbuilder | 100 | 149.0k | 6d ago | SKILL.md |
Frequently asked questions
- How do I install assumption-mapping?
- Run
npx skills add deanpeters/Product-Manager-Skills --skill assumption-mapping. The install tabs above show the steps for each supported agent. - Which AI agents does assumption-mapping work with?
- It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is assumption-mapping safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It declares no license and scores 88/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is assumption-mapping still maintained?
- The repository was last updated about 3 months ago, so assumption-mapping is actively maintained.
Skill content
View source on GitHubname: assumption-mapping argument-hint: "[the idea or plan, and the decision it supports]" description: "List the assumptions behind an idea and rank them by importance and evidence. Use when a plan feels solid and you need to find the beliefs it rests on before you commit." intent: >- Surface the assumptions hiding inside an idea, a plan, or a business case, sort them into desirable, usable, viable, feasible, and ethical risk types, and place each on an importance by evidence map so the team tests the few that can sink the idea first. Output is a ranked list that feeds the de-risking skills. type: component theme: validation-experiments best_for:
- "Finding what an idea silently depends on before building or funding it"
- "Deciding which belief to test first when you cannot test everything"
- "Turning a confident plan into a short list of testable bets"
- "Giving a team a shared picture of what is known versus believed" scenarios:
- "We're about to commit a quarter to this feature. What are we assuming?"
- "Which of these 30 beliefs should we test first?"
- "The business case says 20% attach. What has to be true for that?"
- "Map the assumptions behind our retrofit migration plan" estimated_time: "30-60 min"
Assumption Mapping
Purpose
Find the beliefs your idea silently depends on, rank them, and pick the few worth testing first.
Every plan is a stack of assumptions: customers have this problem, they'll notice our solution, they'll pay for it, we can build it, the channel will carry it. Most are fine. A few can sink the idea, and the dangerous ones are usually the ones nobody wrote down because they felt like facts.
The sentence to keep in your head: you don't need to test everything, you need to test the assumption that would hurt most if it were wrong and that you know least about.
It sits between discovery and de-risking: research-synthesis
tells you what you've learned, this skill tells you what you're still taking on faith, and the Define
skills (pol-probe-advisor,
derisk-measurement-advisor) test it.
Input
Works best with: The idea or plan, and the decision it supports (build it, fund it, launch it).
Also useful: The business case, the research you have, the customer and segment, and who must believe the plan for it to go ahead.
Anything supplied with the invocation itself (text after the skill name, a pasted context dump, or
an appended ARGUMENTS: line) counts as answers already given. Use it and skip whatever it covers;
don't re-ask.
Arriving empty-handed? That works too. The skill opens by asking what the idea is and what has to be true for it to work, and builds the list from your answers. Assumptions it adds itself are tagged as suggestions for you to confirm.
Example invocations:
Assumption map for the Route Optimizer add-on. Business case says 15% attach in year one.What are we assuming in the NFA-500 migration plan? Decision: fund the bracket kit.
Key Concepts
What an Assumption Is
A statement you believe that the plan needs to be true, and that you have not verified. Written as a testable claim, not a worry: "Dispatchers will try the optimizer without a training session," not "adoption might be hard."
Five Kinds of Risk
Sort each assumption by the kind of way the idea could fail:
| Type | The question | Example | |---|---|---| | Desirable | Do customers want it, and does it solve a problem they feel? | Dispatchers feel the rescheduling pain weekly | | Usable | Can they actually use it without help? | A dispatcher can optimize a day in under five minutes | | Viable | Does it work for the business: revenue, cost, channel, legal? | Resellers will sell an add-on for a 15% margin | | Feasible | Can we build and run it? | Routing runs under two seconds at peak load | | Ethical | Could it cause harm or break trust? | Technician location data is acceptable to technicians |
Four of these come from the common product-risk framing (value, usability, viability, feasibility).
The fifth is added here because AI and location-data products fail on it, and nobody writes it
down. See recommendation-canvas and
lean-ux-canvas for related framings.
Generating Them: Ask "What Would Have to Be True?"
Walk the idea as a story, from first contact to payment to long-term use, and at each step ask what has to be true. Do it by role (the user, the buyer, the person who builds it, the partner who sells it) because each sees different risks. Include the boring ones (a team will actually have time to do this; legal will approve). Boring assumptions sink more ideas than clever ones.
Importance by Evidence
Place each assumption on two axes:
- Importance: if this is wrong, does the idea fail? (high = it does)
- Evidence: what do we actually have? (strong = observed or measured; weak = opinion)
| | Strong evidence | Weak evidence | |---|---|---| | High importance | Monitor. You know enough. | Test first. This is the leap of faith. | | Low importance | Ignore | Park it. Revisit if the idea changes. |
The top-right cell is the work. Everything else is either known or doesn't matter yet.
Evidence Is What Happened, Not What Was Said
Rank evidence by how close it is to real behavior: observed behavior and paid usage beat stated
intent, which beats expert opinion, which beats your own conviction. "Everyone I talked to loved it"
is weak evidence unless they paid, scheduled, or did something that cost them. Label what you hold as
Fact, Inference, or Assumption, as in autonomous-investigation.
Pick the Few, and Name What Would Change Your Mind
For each top assumption, state the kill condition: what result would make you change or stop.
Without it, the test only reassures. Then hand the top two or three to a test:
pol-probe-advisor for a proof-of-life probe,
derisk-measurement-advisor for how to measure it.
Facilitation Source of Truth
Use workshop-facilitation as the interaction protocol when
running this conversationally. Assumption mapping works well as a short team workshop: individuals
write assumptions alone first, then the group places them.
Default entry mode: Guided · set in pm-skills.config.yaml (library-wide default or a per-skill override); the user's own choice in the session always wins. On the first turn, offer all three modes, then open in the default without waiting for a pick: Guided asks the first question, Context dump invites the paste, Best guess asks for a one-sentence minimum and drafts with every assumption labeled.
Anti-Patterns (what this is NOT)
- Not a risk register. Risks are worries; assumptions are testable claims the plan needs.
- Not a test plan. It chooses what to test. The Define skills design the test.
- Not a verdict on the idea. It shows what's unproven, not that the idea is wrong.
- Not a list of everything. Thirty assumptions ranked is useful; a hundred is noise.
Application
Use template.md for the scaffold.
Output convention. Artifact first: the assumptions table, the map, the top three with kill
conditions. Commentary short, after. Tag each assumption [FROM YOU], [RESEARCH], or
[AI-SUGGESTED, VALIDATE THIS], and record evidence only if you can name it. Evidence you can't
point to is "none."
If the PM brings a list, critique first. Are they testable claims or worries? Is every risk type represented, or only desirability? Are the boring ones present? Is evidence behavior or opinion? Name gaps; offer rewrites last.
Step 1: Frame the Idea and the Decision
What the idea is, who it is for, and what decision it supports. The decision sets how much evidence is enough.
Step 2: Generate Assumptions
Walk the story by role, asking what must be true. Include boring ones. Write each as a testable claim and tag its risk type.
Step 3: Rate Importance and Evidence
For each, mark importance (does the idea fail if wrong?) and evidence (what do we actually hold, named). Be harsh on evidence: opinion is weak.
Step 4: Map and Choose
Place on the map. Pick the high-importance, weak-evidence assumptions. Choose the top two or three, and write a kill condition for each.
Step 5: Hand Off and Hand Back
Point each top assumption to a test skill. Close with a short summary (supplied, inferred,
unknown) and a numbered ladder, evidence-favored first, labeled (Recommended): probe the top
assumption this week; gather evidence for two more before deciding; run the map as a team workshop;
or stop if a top assumption already looks fatal.
Examples
See examples/sample.md (Fieldlight Route Optimizer, SaaS) and
examples/sample-industrial.md (Northfield NFA-500 migration).
Both show the boring assumption being the one that matters.
Common Pitfalls
Pitfall 1: Only Desirability
Symptom: Every assumption is "customers will want it." Consequence: The idea passes a demand test and dies on cost, channel, or legal. Fix: Walk all five risk types by role.
Pitfall 2: Worries Instead of Claims
Symptom: "Adoption might be hard." Consequence: Nothing to test and no way to know when it's settled. Fix: Restate as a testable claim with a threshold.
Pitfall 3: Opinion Counted as Evidence
Symptom: "Strong evidence: the team is convinced." Consequence: The riskiest belief looks safe. Fix: Evidence is behavior or data you can name. Otherwise it's none.
Pitfall 4: Testing What's Easy
Symptom: The team tests the cheap assumption, not the dangerous one. Consequence: Progress that changes no decision. Fix: Test the high-importance, weak-evidence one first.
Pitfall 5: No Kill Condition
Symptom: The test has no result that would change the plan. Consequence: It only ever reassures. Fix: Write what result stops or changes the idea before running it.
Pitfall 6: Skipping the Boring Ones
Symptom: No assumptions about capacity, legal, support, or sales readiness. Consequence: The idea fails on the item nobody listed. Fix: Ask each role what must be true for them.
References
Skill Files
template.md: assumptions table, map, top three with kill conditionsexamples/sample.md,examples/sample-industrial.md
Related Skills
research-synthesis: upstream; what you've learnedopportunity-solution-tree: where assumptions about solutions get testedpol-probe-advisorandpol-probe: test an assumption cheaplyderisk-measurement-advisor: how to measure itlean-ux-canvasandrecommendation-canvas: related framingsepic-hypothesis: turn the chosen assumption into a hypothesisautonomous-investigation: Fact / Inference / Assumption vocabularyworkshop-facilitation: interaction protocol
External Frameworks
- Assumption mapping, leap-of-faith assumptions: Eric Ries, The Lean Startup; David J. Bland and Alex Osterwalder, Testing Business Ideas.
- Four product risks: Marty Cagan, Inspired.
- Importance by evidence: a widely used 2x2 in lean p
Truncated for display — read the full file on GitHub.
Related Skills
x402
121.0kSet up Browser Use Cloud payments with x402 — pay per request from a crypto wallet (USDC on Base mainnet), no signup or API key.
algorithmic-art
177.9kCreating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems.
pptx
177.9kUse this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx or .potx file (even if the extracted content will be used elsewhere, like in an em…
design
149.0kComprehensive design skill: brand identity, design tokens, UI styling, logo generation (55 styles, Gemini, Atlas Cloud, or MuAPI AI), corporate identity program (50 deliverables, CIP mockups), HTML presentations (Chart.js), banner design (22 styles, social/ads/web/print), icon design (15 styles, SVG…
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
