agent-skills-catalog-maintenance

Detect metadata drift and generate catalog snapshots for agent-skills repositories.

Updated Mar 27, 2026
One-click install
npx skills add https://github.com/gmackie/agent-skills --skill agent-skills-catalog-maintenance-gmackie
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-skills-catalog-maintenance
Source: https://github.com/gmackie/agent-skills/tree/main/skills/agent-skills-catalog-maintenance
Command: npx skills add https://github.com/gmackie/agent-skills --skill agent-skills-catalog-maintenance-gmackie

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires bash, jq, findutils, and includes scripts (resource) components.

What problem does it solve? As skills, tools, and agent definitions are added or migrated in an agent-skills style repo, the discovery surface drifts: SKILL.md files appear without skill.json metadata, helper tools go missing, and README inventories go stale. This Skill keeps the catalog coherent by scanning metadata-backed entries and flagging contract drift. ## Core Features & Use Cases - Metadata Discovery: Scans for skills//skill.json, tools//tool.json, and agents/*/agent-metadata.json while ignoring stray notes and experiments. - Drift Detection: Flags SKILL.md files without metadata, metadata pointing at missing files, missing helper tools, and stale repo-facing inventories. - Snapshot Generation: Runs a bundled bash script that emits a JSON-lines catalog snapshot derived from repo metadata files. - Use Case: After migrating a batch of skills into a repo, ask the agent to update the catalog; it scans metadata, reports drift, generates a snapshot, and updates only the docs that are clearly stale. ## Quick Start Update the catalog after we migrated skills and report any missing metadata or drift.

Frequently Asked Questions about agent-skills-catalog-maintenance

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

FAQPage Schema
How do I keep an agent-skills repo catalog up to date?▼

Scan the repo for skill.json, tool.json, and agent-metadata.json files, report any drift or missing metadata, then run the generate-catalog-snapshot.sh helper to produce a derived inventory. Update README or catalog docs only when they are clearly stale.

How to detect skills missing skill.json metadata?▼

The drift check flags any SKILL.md that exists without a sibling skill.json, metadata pointing at a missing SKILL.md, referenced helper tools that are absent, and agent definitions lacking agent-metadata.json.

What dependencies does the catalog snapshot script need?▼

The snapshot helper requires bash, jq, and findutils. It uses find to locate metadata files and jq to merge each JSON document with its relative source path into a JSON-lines output.

Is the generated catalog snapshot the source of truth?▼

No. The snapshot is a derived artifact. The canonical source of truth remains the repo metadata files themselves, so README text and generated inventories should never be treated as authoritative.

When should I not use this catalog maintenance workflow?▼

Do not use it to invent new skill ideas from scratch; it is designed for maintenance after content is added or migrated. It also ignores temporary experiments and copied artifacts that lack metadata.