Skip to main content
skill CX Codex

023-assumption-analysis — Claude Code skill for assumption analysis

023-assumption-analysis is a skill that identifies and ranks assumptions and unknowns in a framed problem before design, then outlines validation checks.

Apache-2.0 140 installs ★ 465 updated 2026-10-21

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add jabrena/plinth --skill 023-assumption-analysis -a claude-code
Open on GitHub

What 023-assumption-analysis does

Use it after a problem has been framed and its root causes explored. The skill separates claims believed to be true but not yet checked from facts that remain unknown, then ranks both by impact and confidence.

It produces an assumptions list, an unknowns list, and a validation plan that says how and when to check high-impact items with low confidence. Before applying the guidance, the agent must read references/023-assumption-analysis.md. When the problem description is too vague, the skill calls for clarification instead of inventing assumptions or tests.

What it helps with

  • surface hidden beliefs in a problem frame
  • separate unverified claims from unknown facts
  • rank risks by impact and confidence
  • plan checks for high-risk items
  • draft an Assumption Analysis section for a Functional Specification

Example prompts for the agent

  • Before design starts, identify the assumptions and unknowns in this problem.
  • Rank each item by impact and confidence, then propose how and when to validate it.
  • Draft an Assumption Analysis section for this Functional Specification.

How to install 023-assumption-analysis

1Skills CLI (Vercel)

npx skills add jabrena/plinth --skill 023-assumption-analysis -a claude-code

2Manually: put the files in place

~/.claude/skills/023-assumption-analysis/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/023-assumption-analysis/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about 023-assumption-analysis

How do I install 023-assumption-analysis in Claude Code?

Run npx skills add jabrena/plinth --skill 023-assumption-analysis -a claude-code. The skill file is at ~/.claude/skills/023-assumption-analysis/SKILL.md.

What does 023-assumption-analysis do?

It identifies assumptions and unknowns in an explored problem, ranks them by impact and confidence, and prepares a plan to validate them.

What happens when the problem description is vague?

The skill should flag the need for clarification rather than inventing assumptions, unknowns, or validation steps.

Source text

SKILL.md · Apache-2.0
---
name: 023-assumption-analysis
description: Use when a framed and root-caused problem needs its assumptions made explicit before design or planning begins — explicit Assumptions, Unknowns, and a Validation plan. This should trigger when an issue's Assumption Analysis point of view needs evaluation, or when a maintainer directly asks to surface hidden assumptions and unknowns before committing to an approach. Part of Plinth Toolkit
license: Apache-2.0
metadata:
  author: Juan Antonio Breña Moral
  version: 0.19.0
---
# Assumption Analysis

Guide production of explicit Assumptions, a list of Unknowns, and a Validation plan for a problem under exploration. **This is an interactive SKILL**.

**What is covered in this Skill?**

- Surfacing assumptions implicit in the problem frame and root-cause findings
- Distinguishing an assumption (believed true, not yet verified) from an unknown (not yet known either way)
- Ranking assumptions and unknowns by impact and confidence
- Defining a validation plan that names how and when each risky assumption or unknown will be checked
- Feeding assumption and unknown findings into `024-context-mapping` and the remaining Functional Specification lenses

## Constraints

Make assumptions and unknowns explicit before they become undiscussed risk. When this technique is orchestrated by another workflow, the orchestrator owns clarifying-question sequencing; when applied standalone, ask directly.

- **MUST** read `references/023-assumption-analysis.md` before applying Assumption Analysis guidance
- **MUST** state each assumption as a falsifiable claim believed true but not yet verified
- **MUST** distinguish assumptions (believed true) from unknowns (not yet known either way)
- **MUST** rank assumptions and unknowns by impact if wrong and by current confidence
- **MUST** define a validation plan naming how and when each high-impact, low-confidence assumption or unknown will be checked
- **MUST NOT** invent an assumption, unknown, or validation step when the available content is vague or ambiguous; flag the gap for a clarifying question instead

## When to use this skill

- Surface the assumptions behind this problem
- List the unknowns for this issue
- Build a validation plan for these assumptions
- Apply assumption analysis before design begins
- Draft the Assumption Analysis section of a Functional Specification

## Workflow

1. **Read the Reference**

Read `references/023-assumption-analysis.md`, then review the problem frame and root-cause findings for implicit beliefs.

2. **Surface Explicit Assumptions**

State each assumption as a falsifiable claim believed true but not yet verified.

3. **List Unknowns**

List facts that are not yet known either way, distinct from assumptions.

4. **Rank by Impact and Confidence**

Rank assumptions and unknowns by impact if wrong and by current confidence, prioritizing high-impact, low-confidence items.

5. **Define the Validation Plan**

Name how and when each high-priority assumption or unknown will be validated.

6. **Report the Assumption Analysis**

Report the Assumptions, Unknowns, and Validation plan, and flag any item left open pending a clarifying answer.

## Reference

For detailed guidance, examples, and constraints, see [references/023-assumption-analysis.md](references/023-assumption-analysis.md).

© Juan Antonio Breña Moral · Apache-2.0 · jabrena/plinth

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set123

The whole set →
812-regulations-eu-product-liability-directive 812-regulations-eu-product-liability-directive turns EU Product Liability Directive 2024/2853 checks into engineering artefacts: scope inventory, hazardous-scenario tests, update history and user warnings for Java and AI products. CX Codex +1 165installs 810-regulations-eu-mifid-ii 810-regulations-eu-mifid-ii reviews Java enterprise applications for MiFID II investment service controls and audit evidence prior to release. CX Codex +1 165installs 022-root-cause-analysis 022-root-cause-analysis — a skill for Claude Code, Codex, Cursor, GitHub Copilot, and Gemini CLI that runs structured root cause analysis using Five Whys, Fishbone (Ishikawa), and Current Reality Tree. CX Codex +1 144installs 025-quality-attribute-discovery The 025-quality-attribute-discovery skill finds and ranks a solution’s quality attributes before architecture work begins. It ties each candidate to evidence from the problem and its context. CX Codex +1 143installs 024-context-mapping 024-context-mapping is a Claude Code skill that maps existing systems, integrations, owners, and external dependencies before design when system boundaries are undefined. CX Codex +1 142installs 021-problem-framing 021-problem-framing structures problem statements, current and desired states, stakeholders, and success criteria before engineers jump into implementation. CX Codex +1 139installs 059-design-atdd 059-design-atdd is a Claude Code skill that checks whether an OpenSpec change’s goal, acceptance criteria, and implementation or verification tasks agree. It produces an evidence-backed report of gaps and mismatches. CX Codex +1 135installs 058-design-bdd 058-design-bdd — a skill that takes Java teams through Behavior-Driven Development as collaborative discovery practice. Works with trusted behavior facts, concrete examples, and scenarios in shared domain language. CX Codex +1 135installs

Project sponsors

Companies that keep this analytics open