Assumption Auditor: Find the Hidden Risks in Any Plan

    1

    Surfaces every hidden assumption in a plan, PRD, pitch, architecture doc or essay, ranks each by risk, and gives a concrete way to validate it before you commit weeks to the wrong thing.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Assumption Auditor: Find the Hidden Risks in Any Plan

    Example session with this skill installed

    Audit the assumptions in this plan: We're building a project management tool specifically for freelancers. Unlike Asana or Trello which are built for teams, ours will focus on solo workflows — client tracking, invoice reminders, and a Kanban board. We'll charge $12/month. Target market is the 59 million freelancers in the US. We'll grow through content marketing and launch in Q2, aiming to break even by Q4.

    • Read your context and instructions
    • Compiled the assumption auditor

    Assumption audit: "project management tool for freelancers..."

    Total assumptions found: 12
    High-risk (if wrong, plan fails): 5
    Medium-risk (if wrong, plan weakens): 4
    Low-risk (if wrong, plan adjusts): 3


    High-risk assumptions

    1. Freelancers are meaningfully underserved by current tools and actively want a replacement.

      • Type: Foundational
      • Where it appears: "Unlike Asana or Trello which are built for teams"
      • Why it matters: If freelancers are already satisfied adapting team tools, there is no pull for a new product.
      • How to validate: Interview 20 freelancers. Ask what tool they use today and what it would take to make them quit it.
    2. The addressable market is large enough and concentrated enough to grow through content.

      • Type: Market/User
      • Where it appears: "59 million freelancers in the US"
      • Why it matters: Most of those 59M are gig workers with no PM software need. The real ICP is likely designers, developers, writers, consultants — a much smaller number.
      • How to validate: Narrow ICP to one freelancer type and size that segment specifically.
    3. Freelancers in the target segment will pay $12/month for this.

      • Type: Market/User
      • Where it appears: "We'll charge $12/month"
      • Why it matters: No validation cited. Could be too high, too low, or the wrong model entirely.
      • How to validate: Ask 30 target users directly: "Would you pay $12/month? Show me your current tool spend."
    4. Content marketing will generate enough growth to break even by Q4.

      • Type: Organizational
      • Where it appears: "We'll grow through content marketing and... break even by Q4"
      • Why it matters: Content marketing typically takes 12–18 months to compound. Q4 break-even implies fast organic traction that content rarely delivers.
      • How to validate: Model it — what monthly signups are needed, what conversion rate, what content volume?
    5. No well-funded competitor already targets this niche.

      • Type: Market/User
      • Where it appears: Implied by the entire premise
      • Why it matters: Bonsai, HoneyBook, and Dubsado already target freelancers specifically.
      • How to validate: Full competitive audit before writing another line of code.

    Medium-risk assumptions

    1. Bundling invoicing, client tracking, and Kanban into one tool is what freelancers want.
      • How to validate: Ask which of the three they'd pay for standalone and which they'd migrate.
    2. A Q2 launch is achievable given current team capacity.
      • How to validate: Break into tasks, estimate hours, stress-test Q2 date against actual capacity.
    3. Freelancers prefer SaaS subscription over one-time payment.
    4. Kanban is the right interaction model for solo workers.

    Low-risk assumptions

    • $12/month is preferable to $8 or $19 as a price point
    • US market should be targeted before international
    • The product will be web-first rather than mobile-first

    The one assumption most likely to be wrong

    "Built for freelancers" is the reason they don't love current tools. But freelancers use Notion, Trello, and Toggl extensively and successfully. The plan treats positioning as product-market fit. Those are different things.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Most plans fail not because of bad execution, but because of invisible assumptions that were never written down or challenged. When you write a plan you implicitly treat dozens of statements as true: that the problem is real, that people will pay, that you have enough time, that the technology will cooperate. You don't list them because they feel obvious. That's exactly what makes them dangerous.

    This skill makes your agent a critical-thinking auditor. It digs out every assumption in what you share, classifies it, rates the risk, and gives you a concrete way to test each one, so you can find out cheaply before you commit.

    What you get

    • Every assumption stated as a plain, testable claim ("Users will pay $50/month"), never as a vague question.
    • Four types covered: Foundational (the problem exists and is worth solving), Market/User (people will want, pay, behave this way), Technical (it will perform, scale, integrate), and Organizational (we have the skills, time, resources and buy-in).
    • Risk-ranked: high (if wrong, the plan fails), medium (the plan weakens), low (the plan adjusts).
    • For each high-risk assumption: where it appears in your text, why it matters, and a concrete validation step in under 20 words.
    • The one assumption most likely to be wrong: the one you are most confident about and least aware of.

    Works on

    • Business plans, pitch decks and go-to-market plans
    • Architecture and technical design docs
    • Product requirements documents (PRDs) and roadmaps
    • Essays and arguments
    • Personal plans: career moves, side projects, big decisions

    How it behaves

    • It surfaces assumptions; it doesn't judge your plan as good or bad. You decide what to do with them.
    • It assumes good intent: the point is to make the invisible visible, not to criticize.
    • It always finds at least one foundational assumption, and it digs harder on short documents, which hide more assumptions per word.

    Prompt-only: one SKILL.md, no scripts, no API keys, nothing to install. Includes three worked examples (SaaS plan, architecture migration, personal plan).

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Product Management.

    It works from what you share: the skill doesn't research market data or competitors itself, so its validation steps tell you what evidence to go get. It surfaces assumptions and rates their risk; it doesn't tell you whether to go ahead. Very long documents are best audited section by section.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 5 days ago

    • Free with an Agensi account

    Requires

    Nothing to install.

    Permissions

    Read Files

    Read Files only: if you point the agent at a document instead of pasting it, the agent reads that file. The skill has no scripts, runs nothing, writes nothing and makes no network calls.

    Listed5 days ago

    What's inside

    Frequently Asked Questions

    A general critique mixes opinions, style notes and vague questions. This skill looks for one thing, the statements your plan treats as true without proof, and states each as a testable claim with its type, where it appears, what breaks if it's wrong, and a concrete way to validate it. The output is a ranked checklist you can act on, ending with the single assumption most likely to be wrong.

    Anything where being wrong matters: business plans, pitch decks, PRDs, roadmaps, architecture and technical design docs, strategy proposals, essays and arguments, and personal plans like a career move or side project. Short documents work well too; the skill digs harder on them because they hide more assumptions per word.

    No. It assumes you have good reasons for each assumption and never judges the plan as a whole. Its job is to make the invisible visible and rank the risk; you decide what to do with it.

    Foundational: the problem exists and is worth solving (rarely stated because it feels obvious). Market/User: people will want, pay for or behave a certain way (the most often wrong). Technical: a system will perform, scale or integrate as expected. Organizational: you have the skills, time, resources and buy-in (the most often invisible).

    New in Product Management

    Free