verify

Run structured audits against NotebookLM to validate ingested knowledge.

16|9|Updated Apr 4, 2026
One-click install
npx skills add https://github.com/sonthanh/brain-os-plugin --skill verify-sonthanh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: verify
Source: https://github.com/sonthanh/brain-os-plugin/tree/main/skills/verify
Command: npx skills add https://github.com/sonthanh/brain-os-plugin --skill verify-sonthanh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Inaccurate ingestion of self-learned knowledge can lead to degraded decisions; Knowledge Verify provides an automated audit to catch and flag errors before absorption.

Core Features & Use Cases

  • Generates 50 fresh questions across topics, cross-cutting areas, and adversarial prompts to validate ingested notes.
  • Uses NotebookLM as an objective judge and stores audit flags and results locally for governance.
  • Supports CLI controls to run verification, check status, and assign audit flags per knowledge book.

Quick Start

Run verify on a selected book to generate 50 fresh questions, evaluate with NotebookLM, and update the audited flag.

Frequently Asked Questions about verify

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

FAQPage Schema
How do I audit notes in an Obsidian vault for ingestion errors?

Auditing notes in an Obsidian vault requires running a structured verification against NotebookLM to catch validation gaps. The process generates 50 fresh questions across topics to evaluate ingested notes and stores audit results in the book's _validation directory.

What's the best way to validate knowledge ingested into NotebookLM?

Validating knowledge ingested into NotebookLM is best handled by generating cross-cutting and adversarial questions to test the source material. This automated audit uses NotebookLM as an objective judge to flag validation gaps before the knowledge is absorbed.

Can I run automated verification checks across multiple books in my vault?

Automated verification checks can run across multiple vault books using fuzzy matching to select targets. CLI controls allow you to execute the audit, check verification status, and assign audit flags per knowledge book individually.

Does knowledge verification support CLI commands for checking audit status?

Knowledge verification supports CLI commands to run the audit, check status, and set audit flags. It outputs the validation results locally into the book's _validation directory and updates the audited flag accordingly for governance.

Why do I need an objective judge like NotebookLM for self-learned knowledge validation?

An objective judge like NotebookLM is needed for self-learned knowledge validation because inaccurate ingestion can lead to degraded decisions. Generating adversarial prompts tests the ingested notes against an impartial source to catch and flag errors before absorption.