auditor

Audit discovery outputs across five weighted dimensions and generate a convergent validation report.

Updated Apr 10, 2026

Search Tags:#discovery #evaluation #auditing #reporting #quality #gates #convergent

Ready to run

npx skills add https://github.com/fabioabr/ia-ws --skill auditor-fabioabr

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: auditor
Source: https://github.com/fabioabr/ia-ws/tree/main/projects/discovery-to-go/base/behavior/skills/auditor
Command: npx skills add https://github.com/fabioabr/ia-ws --skill auditor-fabioabr

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Audits discovery results to enforce quality gates across all blocks and surfaces actionable findings in a formal report.

Core Features & Use Cases

Evaluates the 8 discovery blocks across 5 weighted dimensions (completeness, grounding, coherence, depth, neutrality).
Produces a convergent validation report (2.1-convergent-validation.md) with a score and verdict.
Runs in parallel with the 10th-man gate, enabling independent quality checks.

Quick Start

Run the auditor on the active iteration to generate the 2.1-convergent-validation.md report.

Frequently Asked Questions about auditor

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

FAQPage Schema

How do I automate discovery quality audits for Phase 2 challenge iterations?▼

You can automate discovery quality audits by running the auditor Skill on your active iteration, which analyzes five drafts, interview logs, and briefs to generate a formal convergent validation report with a pass or fail verdict.

What dimensions are used to evaluate discovery output quality?▼

Discovery output quality is evaluated across five weighted dimensions: completeness, grounding, coherence, depth, and neutrality, ensuring all eight discovery blocks meet predefined quality gates before receiving a formal score.

How do I generate a convergent validation report for discovery drafts?▼

To generate a convergent validation report, apply the auditor to your Phase 2 challenge iteration, which outputs the 2.1-convergent-validation.md file containing the formal score, verdict, and actionable findings from the quality audit.

Can I run independent quality checks alongside the 10th-man gate process?▼

Yes, you can run independent quality checks alongside the 10th-man gate process, as the auditor operates in parallel to enforce quality gates across all blocks without interference from the independent review mechanism.

What inputs do I need to evaluate discovery outputs and determine a pass or fail verdict?▼

You need five discovery drafts, interview logs, and briefs as inputs for the auditor to apply its scoring, thresholds, and conditional checks, ultimately determining a formal pass or fail verdict for the iteration.

When should I use an automated discovery auditor instead of manual evaluation?▼

You should use an automated discovery auditor when you need to consistently enforce quality gates across all eight discovery blocks, requiring a formal score, actionable findings, and a standardized convergent validation report for Phase 2 iterations.