Why do I need to verify plans if the AI wrote them?

AI agents can silently drop up to 14% of requirements during the planning phase. This skill makes those gaps visible and actionable before you waste time on incomplete execution.

Can it detect mismatched status codes or file limits?

Yes. Phase 6 of the algorithm performs a Mutation Scan that compares numbers, HTTP codes, and thresholds character-by-character to ensure the plan hasn't drifted from the spec.

What documents does this skill require?

It typically looks for markdown specs and architecture documents in your project's _docs folder, but can also be pointed to specific files via arguments.

How does this skill prevent AI hallucinations?

It uses strict anti-hallucination rules that forbid 'implied' coverage. Every requirement must be explicitly matched to a specific task and behavior with exact value parity.

Plan Coverage Verifier

Name: Plan Coverage Verifier
Author: fiatkongen

byfiatkongen

0•

개발자 도구

Validates implementation plans against technical specifications and architecture documents to prevent requirement drops and configuration mutations.

The Plan Coverage Verifier is a high-precision quality assurance skill designed to bridge the gap between planning and execution. It systematically extracts requirements, user flows, and architectural contracts from your documentation and cross-references them against your implementation plans. By enforcing strict anti-hallucination rules and performing character-by-character value comparisons, it identifies missing requirements, partial implementations, value mutations, and orphan tasks. This ensures that before any code is written, your implementation strategy is 100% aligned with your original design specifications.

주요 기능

01Orphan task detection for implementation steps not backed by architectural contracts

02Multi-phase forward and backward checking to ensure total plan completeness

03Specific, actionable fix instructions for all identified gaps and mismatches

04Automated requirement and contract extraction from specs and architecture docs

05Strict mutation scanning for exact matching of status codes, limits, and thresholds

060 GitHub stars

사용 사례

01Verifying a complex feature plan before starting subagent-driven development

02Audit-logging requirement coverage for high-stakes enterprise implementations

03Ensuring API implementation plans perfectly match defined REST/GraphQL contracts

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add fiatkongen/saurun-marketplace verify-plan-coverage

For use in Claude.ai and ChatGPT

Download Skill