How does this skill handle low-confidence assumptions?

The skill follows a 'Stop and Ask' process, requiring the model to present alternatives to the user for clarification rather than making a commitment to a specific interpretation.

Can this skill improve my documentation and testing?

Yes, by identifying assumptions, it provides a clear roadmap for what needs to be documented in comments, enforced via types (like TypeScript), or covered by unit tests.

Why is surfacing assumptions important for AI coding?

AI models often make silent interpretations of requirements; surfacing assumptions forces these to be explicit, preventing bugs caused by incorrect 'guesses' made by the model during code generation.

When is the best time to trigger the Surface Assumptions skill?

It is most effective when used during the planning phase before writing new code, or during a code review phase to audit the logic of existing snippets.

Surface Assumptions

Name: Surface Assumptions
Author: rahulsub

byrahulsub

•

Security & Testing

Identifies and validates hidden technical assumptions to prevent logic errors and edge-case bugs before implementation.

The Surface Assumptions skill addresses the most common category of AI coding errors: models making incorrect assumptions and proceeding without verification. By forcing an explicit audit of input shapes, state requirements, environmental dependencies, and business logic, this skill ensures that Claude clarifies ambiguous requirements rather than guessing. It provides a structured framework for categorizing assumptions by confidence and impact, leading to more robust, production-ready code with fewer runtime failures and better-handled edge cases.

Key Features

011 GitHub stars

02Standardized output format for documenting validated constraints

03Runtime validation check generation for critical assumptions

04Explicit listing of input, state, and environment assumptions

05Structured 'Stop and Ask' workflow for low-confidence logic

06Risk categorization based on confidence and impact levels

Use Cases

01Defining edge-case handling and validation logic for external API integrations

02Peer review of AI-generated code to catch hidden hallucinations or incorrect interpretations

03Pre-implementation planning for complex features to ensure alignment with business rules

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rahulsub/code-practices-plugin surface-assumptions

For use in Claude.ai and ChatGPT

Download Skill