gsd-verify

Verify phase goals by tracing requirements to code, tests, and documentation.

7|2|Updated Jun 30, 2025
One-click install
npx skills add https://github.com/hq-opensource/predictive-control --skill gsd-verify
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gsd-verify
Source: https://github.com/hq-opensource/predictive-control/tree/main/.agent/skills/gsd-verify
Command: npx skills add https://github.com/hq-opensource/predictive-control --skill gsd-verify

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill ensures that implemented features meet project requirements by performing goal-backward verification, orchestrating UAT, and producing structured verification documentation to prevent false positives of completion.

Core Features & Use Cases

  • Goal-Backward Verification: Validate that built features actually achieve stated goals by tracing requirements to code and artifacts.
  • User Acceptance Testing (UAT): Structure and document user-facing tests to confirm real-world usability and alignment with phase goals.
  • Verification Report Generation: Produce a VERIFICATION.md report that records verdicts, evidence, gaps, and recommended fixes.
  • Subagent Integration: Reference internal subagents (verifier, integration-checker) to perform modular quality checks.

Quick Start

Run the gsd-verify skill to audit a phase and generate its VERIFICATION.md report.

Frequently Asked Questions about gsd-verify

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How to verify implemented features meet phase requirements and prevent false completion positives?

Goal-backward verification validates implemented features against stated phase goals by tracing requirements to code, tests, and documentation, orchestrating UAT, and generating a VERIFICATION.md report with auditable evidence to prevent false completion positives.

What is goal-backward verification for software engineering projects?

Goal-backward verification is a validation process that confirms built features actually achieve stated project goals by mapping requirements backward to code and artifacts, structuring UAT, and recording verdicts, gaps, and recommended fixes in a VERIFICATION.md report.

How do I map requirements to code and tests across development phases?

You map requirements to code and tests by running verification workflows that export and wire checks, orchestrate user acceptance testing, and generate structured documentation providing auditable evidence of phase goal alignment and requirement fulfillment.

Does this verification workflow require external dependencies or subagent modules?

The verification workflow relies on internal subagents like verifier and integration-checker to perform modular quality checks, and requires no external dependencies to map requirements, orchestrate UAT, and produce the final VERIFICATION.md evidence report.

What's the best way to generate auditable UAT artifacts for codebase verification?

Generating auditable UAT artifacts is best achieved by structuring user-facing tests to confirm real-world usability, wiring checks to codebase features, and documenting verdicts, evidence, and gaps within a structured VERIFICATION.md report.

When do I need a VERIFICATION.md report for e2e and integration testing?

You need a VERIFICATION.md report when completing a development phase to provide auditable evidence that e2e and integration testing align with requirements, recording validation verdicts, identified gaps, and recommended fixes to guarantee goal-aligned verification.