skills.rest
zivtech
zivtechOFFICIAL
Explore all skills by this author →
WebsiteGitHub

review-skill

Correlate observation logs with verified evaluation artifacts to diagnose skill performance issues.

Updated Mar 5, 2026
Search Tags:#debugging#optimization#audit#evaluation#maintenance#observability
One-click install
npx skills add https://github.com/zivtech/joyus-desktop --skill review-skill-zivtech
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: review-skill
Source: https://github.com/zivtech/joyus-desktop/tree/main/.claude/skills/review-skill
Command: npx skills add https://github.com/zivtech/joyus-desktop --skill review-skill-zivtech

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python3, and includes scripts (resource) components.

What problem does it solve?

This Skill addresses the silent degradation of AI skills by automating the inspection of performance logs and proposing evidence-based improvements to ensure long-term reliability.

Core Features & Use Cases

  • Health Diagnosis: Analyzes observation logs and evaluation results to identify patterns of failure or performance drift.
  • Evidence-Based Amendment: Generates targeted, minimal prompt changes to fix identified issues based on verified run data.
  • Use Case: Use this tool to perform a quarterly health check on your active skills to detect if model updates or codebase changes have caused a drop in accuracy or tool-use reliability.

Quick Start

Invoke the review skill by providing the name of the skill you wish to analyze and optionally specifying a focus area like routing or instructions.

Frequently Asked Questions about review-skill

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

FAQPage Schema
How do I diagnose performance drift in AI agent workflows?▼

Skill health audits diagnose performance drift in agentic workflows by correlating observation logs with verified run-manifest-v1 evaluation artifacts. This identifies failure patterns and operational issues by grounding diagnostic data in schema-valid execution history.

How do I fix silent degradation in AI skills after a model update?▼

Fix silent degradation in AI skills by generating evidence-based, targeted prompt amendments. The system proposes minimal changes to instructions based on verified run data to restore long-term reliability and tool-use accuracy after updates.

Do I need python3 to run automated skill health checks?▼

Yes, python3 is a required dependency to run automated skill health checks. The environment uses python3 to execute scripts that analyze observation logs and generate structured health reports for agentic workflows.

What is the best way to audit agentic workflow performance using execution history?▼

The best way to audit agentic workflow performance is by validating schema-valid run-manifest-v1 artifacts. This ensures all diagnostic data is grounded in verified execution history, enabling structured health reporting and accurate issue detection.

Can I target a specific focus area like routing when evaluating skill health?▼

Yes, you can target a specific focus area like routing or instructions when evaluating skill health. Specifying a focus area narrows the analysis of observation logs to generate targeted prompt amendment proposals for that exact operational component.

Related Skills

Explore More Software Engineering Skills→
fmschulzfmschulz
7

bio-workflow-methods-docwriter

Generate METHODS.md and run_manifest.yaml from Nextflow, Snakemake, and CWL workflow artifacts.

Community
Advanced
glfruitglfruit

autoresearch

Automate OpenClaw skill optimization through iterative sub-agent runs and prompt mutations.

Community
Advanced
Aniket-a14Aniket-a14
1

senior-data-engineer

Design production-grade data pipelines with Python, Spark, Airflow, dbt, and Kafka.

Community
Advanced
tae2089tae2089
2

agent-team-ops

Inspect runs, messages, and events and finalize runs without executing agents.

Community
Intermediate