prune-skill

Audit agent skills by verifying claims against cited sources and generating prune reports.

3|Updated Apr 5, 2026
One-click install
npx skills add https://github.com/dvy1987/agent-loom --skill prune-skill
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: prune-skill
Source: https://github.com/dvy1987/agent-loom/tree/main/.agents/skills/prune-skill
Command: npx skills add https://github.com/dvy1987/agent-loom --skill prune-skill

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill enables evidence-based pruning of outdated or incorrect content within agent skills, ensuring the library stays current, trustworthy, and aligned with updated research and practices.

Core Features & Use Cases

  • Formal citation audits for each claim and source within skills.
  • Obsolescence checks against known model evolutions and recency windows.
  • Prune log generation and post-prune validation to preserve integrity and traceability.

Quick Start

Run the prune workflow on a target skill to generate a prune log with recommended updates.

Frequently Asked Questions about prune-skill

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

FAQPage Schema
How do I audit agent skills for outdated or incorrect content?

Auditing agent skills involves identifying model-specific or time-sensitive claims and verifying each against cited sources. The tool evaluates techniques against current model evolution and trusted references to produce a structured prune report with corrections and flagged items.

What is a prune report and how does it help with content governance?

A prune report is a structured output containing a prune log, corrections, and flagged items for content governance. It provides traceable updates that keep the skill library current, trustworthy, and aligned with updated research and practices.

How do I verify citations and references in my agent skills?

Citation verification is performed by checking each claim and source within the skill against trusted references and known model evolutions. The tool assesses recency windows to determine whether the citation remains valid or requires replacement.

What's the best way to identify obsolete techniques in AI agent skills?

The best way to identify obsolete techniques is to evaluate each behavior against current model evolution and recency windows. This process validates techniques against current research and trusted references to determine obsolescence or replacement needs.

Do I need any dependencies to run the prune workflow on a target skill?

No dependencies are required to run the prune workflow. The tool operates independently to audit the target skill and generate a prune log with recommended updates based on repository audit standards.

How do I validate updated skills against repository audit standards?

Post-prune validation is performed automatically after generating the prune report. It checks the updated skill against the repository's audit standards to preserve integrity and ensure compliance with governance requirements.