skills.rest
ddobrin
ddobrinCOMMUNITY
Explore all skills by this author →
Website𝕏@ddobrinGitHub

geap-spec-validator

Detect ambiguous and incomplete requirements in drafted software specifications via Vertex AI review.

8|Updated Jun 18, 2026
Search Tags:#claude#requirements analysis#gemini#specification validation#adversarial review#vertex ai
One-click install
npx skills add https://github.com/ddobrin/plan-skills --skill geap-spec-validator
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: geap-spec-validator
Source: https://github.com/ddobrin/plan-skills/tree/main/plugins/plan/skills/geap-spec-validator
Command: npx skills add https://github.com/ddobrin/plan-skills --skill geap-spec-validator

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill exposes ambiguity, logical gaps, incomplete requirements, and untestable acceptance criteria before a software specification reaches planning or implementation. It provides an independent review perspective by running adversarial skeptics on remote Vertex AI foundation models.

Core Features & Use Cases

  • Adversarial Spec Review: Runs three focused skeptics covering ambiguity and malicious compliance, logic and boundaries, and completeness and testability.
  • Model Diversity: Supports configurable Gemini and Claude models through Vertex AI for independent external perspectives.
  • Programmatic Quorum: Combines skeptic votes with a synthesis model and confirms findings only when at least two of four possible votes agree.
  • Audit Reports: Writes validation findings, verdicts, and recommended tightenings to the appropriate milestone review directory.
  • Use Case: Before planning a new feature, submit its specification to the remote panel to identify underspecified behavior, boundary-case failures, and missing verification criteria.

Quick Start

Use the geap-spec-validator skill to run a remote Vertex AI adversarial review on the specified specification file and report the confirmed findings.

Frequently Asked Questions about geap-spec-validator

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

FAQPage Schema
How do I validate a software specification for ambiguity and logical gaps before implementation?▼

Specification validation detects ambiguity, logical gaps, incomplete requirements, and untestable criteria before implementation planning. An adversarial review runs three focused skeptics covering ambiguity, logic, and completeness to expose underspecified behavior.

Can I use Gemini or Claude models on Vertex AI to review my feature specs?▼

Yes, you can use configurable Gemini and Claude models through Vertex AI for independent external perspectives. These remote foundation models run as adversarial skeptics to challenge your drafted feature, bug-fix, or refactoring specifications.

What's the best way to run an adversarial review on untestable acceptance criteria?▼

An adversarial review using a quorum-based reporting system combines three skeptic votes with a synthesis model. It confirms findings only when at least two of four possible votes agree, ensuring reliable detection of untestable acceptance criteria.

Do I need Google Cloud authentication to run remote specification validation?▼

Yes, remote specification validation requires Google Cloud authentication, a configured GCP project, and Python HTTP with Google authentication libraries. These prerequisites enable the skill to run remote Vertex AI foundation models.

How does quorum-based reporting work for specification validation findings?▼

Quorum-based reporting combines votes from three focused skeptics and a synthesis model, confirming validation findings only when at least two of four possible votes agree. This programmatic quorum ensures audit reports contain verified verdicts and recommended tightenings.

When should I not use remote AI models for specification validation?▼

Remote AI specification validation should not be used when independent external perspectives are inappropriate or when Google Cloud authentication, a configured GCP project, and Python libraries cannot be provisioned to run the Vertex AI models.

Related Skills

Explore More Software Engineering Skills→
heyu-aiheyu-ai

verify-gemini-models

Verify Gemini model availability and functionality via REST and WebSocket calls.

Official
Advanced
dobrinz123dobrinz123

gsd-milestone-summary

Generate milestone summaries from .planning artifacts into seven-section reports.

Community
Advanced
ajianazajianaz
4

task-generator

Generate structured task lists with parent and sub-tasks from specs.

Community
Basic
ddobrinddobrin
8

geap-interactions-spec-validator

Identify ambiguity and unverifiable acceptance criteria in software specifications via remote skeptic reviews.

Community
Advanced