skill-5b-semantic-edge-verification

Verify strict edges and unmatched concerns in AI reviewer match graphs.

1|Updated Apr 9, 2026
One-click install
npx skills add https://github.com/jinming99/reviewer-under-review --skill skill-5b-semantic-edge-verification
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-5b-semantic-edge-verification
Source: https://github.com/jinming99/reviewer-under-review/tree/main/.claude/skills/skill-5b-semantic-edge-verification
Command: npx skills add https://github.com/jinming99/reviewer-under-review --skill skill-5b-semantic-edge-verification

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Semantically verifies all strict edges and unmatched concerns by reading the audit worksheet produced by Skill 5a, and returns structured override judgments to feed back into the alignment metrics pipeline.

Core Features & Use Cases

  • Reads the pre-generated audit worksheet and related context (match graphs, official and agentic concern sheets) to validate edge quality.
  • Produces structured override judgments for edges and severity alignments, and exports an overrides YAML for downstream metric computation.
  • Supports documentation of missed matches and phantom checks, enabling traceable audit trails and reproducible recalibration.

Quick Start

Run the semantic audit workflow to generate worksheets, then run the verification to produce the semantic_overrides.yaml.

Frequently Asked Questions about skill-5b-semantic-edge-verification

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

FAQPage Schema
How do I semantically verify unmatched concerns in AI reviewer match graphs?

To semantically verify unmatched concerns, you audit strict edges by cross-referencing match graphs with official and agentic concern sheets to validate edge quality and produce structured override judgments.

How do I generate override judgments for strict edges in a calibration workflow?

You generate override judgments by reading pre-generated audit worksheets and related context to validate edge quality, then exporting an overrides YAML file for downstream metric computation and recalibration.

What is semantic edge verification for concern alignment?

Semantic edge verification is the process of auditing strict edges and unmatched concerns across papers to validate alignment quality, returning structured override judgments for metric recomputation.

Do I need a pre-generated audit worksheet to run semantic edge verification?

Yes, semantic edge verification requires a pre-generated audit worksheet produced by the prior semantic audit workflow, along with match graphs and related concern sheets to cross-reference edge quality.

What format does semantic edge verification output for metric recomputation?

Semantic edge verification outputs a semantic_overrides.yaml file for metric recomputation, along with a verification_summary report documenting missed matches and phantom checks for traceable audit trails.

When should I not use automated semantic edge verification for alignment metrics?

Automated semantic edge verification is not suitable without the initial audit worksheet and related context sheets, as it depends on cross-referencing existing match graphs to validate edge quality across papers.