Skill Health

Audit per-skill quality metrics and run data to classify health and prepare notifications.

Updated Jun 3, 2026
One-click install
npx skills add https://github.com/swarm-ai-research/aeon --skill skill-health-swarm-ai-research
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: Skill Health
Source: https://github.com/swarm-ai-research/aeon/tree/main/skills/skill-health
Command: npx skills add https://github.com/swarm-ai-research/aeon --skill skill-health-swarm-ai-research

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires gh, jq.

What problem does it solve?

Audit skill quality metrics, detect API degradation, file issues for new failures and resolve them when skills recover, and notify only when fleet health state actually changes.

Core Features & Use Cases

  • Monitor memory/cron-state.json for per-skill quality metrics and detect regressions.
  • Read memory/skill-health/*.json and memory/issues/INDEX.md to classify health, track open issues, and deduplicate notifications.
  • Automatically file or resolve issues when skills fail or recover, and notify only on state changes.

Quick Start

Run the skill-health audit to evaluate all enabled skills and file or resolve issues automatically.

Frequently Asked Questions about Skill Health

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

FAQPage Schema
How do I monitor skill reliability and detect API degradation automatically?

You monitor skill reliability by running an audit that analyzes per-skill quality metrics from cron-state.json and recent run data to surface API degradation and failures across enabled skills.

How do I auto-resolve GitHub issues when skills recover?

To auto-resolve issues, the audit reads memory/skill-health/*.json and memory/issues/INDEX.md to track open issues, automatically resolving them in GitHub when skills recover from failures.

Does the skill health audit require jq and gh dependencies?

Yes, the skill health audit requires both jq and gh dependencies to parse JSON state files and automatically file or resolve issues via the GitHub CLI.

What is the best way to deduplicate monitoring notifications for enabled skills?

The best way to deduplicate notifications is to classify health states and trigger alerts only when fleet health actually changes, preventing duplicate warnings for ongoing issues.

What happens if there is no run data when auditing skill health?

When no run data exists, the audit includes built-in safeguards for no-data scenarios, producing a structured output that classifies health states and provides remediation guidance without failing.