Sample Output
A Complete Picture of Your Skill Library
Here's what the generated report looks like:
As your agent's skill library grows, so does the chaos. Duplicate skills, bloated packages, unused tools — you deserve visibility. Skill Inspector automatically audits, scores, and recommends improvements for your entire skill ecosystem.
Every agent accumulates skills over time. Without governance, the library becomes:
Overall score from 0–100 across Structure, Maintainability, Reusability, and Duplication dimensions.
Finds overlapping skills using semantic analysis and cosine similarity clustering.
Identifies monolithic packages, reference bloat, and category concentration risks.
Understands skill distribution across categories, asset types, and package complexity.
Tracks which skills are actually used, unused, and high-cost candidates for archival.
Generates actionable suggestions: split, merge, reduce, or remove based on your data.
Here's what the generated report looks like:
Hermes Agent Cloud comes with Skill Inspector pre-installed. Simply tell your agent
skill health check
and the agent will automatically run the inspection, generate a report, and provide recommendations — all within your Telegram chat.
Get the latest version directly from GitHub:
pip install --upgrade git+https://github.com/johnsonbuilds/skill-inspector.git
Scan your skills and generate a comprehensive report:
# Scan default data directory (/opt/data/skills/)
skill-inspector scan-packages
# Custom directory and output
skill-inspector scan-packages --data-dir /path/to/data --output report.md
Open the generated report.md, follow governance recommendations, and keep your skill library healthy.
Pro tip: For large skill libraries, run scans in the background to avoid blocking your terminal:
skill-inspector scan-packages && echo "Done!"
Full scan with package analysis, duplicate detection, and governance recommendations. Generates report.md.
skill-inspector scan-packages \
--data-dir /opt/data \
--output report.md \
--duplicate-threshold 0.82
Quick health report focusing on structure, maintainability, and duplication scores.
skill-inspector health \
--data-dir /opt/data \
--output health.md
Don't want to manage servers, dependencies, or configurations? Hermes Agent Cloud gives you everything out of the box — including Skill Inspector, pre-installed and ready to use.
Cloud-hosted. No VPS, no Docker, no maintenance.
Chat with your agent. Ask for skill health checks naturally.
Sign up, connect Telegram, start using skills immediately.
Your agent runs 24/7. Skills are always available.
Skill Inspector is open source and hosted on GitHub. We welcome contributions, issues, and feature requests. Join the community and help shape the future of agent skill governance.
Skill Inspector is an automated governance tool for Hermes Agent skill libraries. It scans your installed skills, detects duplicates, calculates health scores, analyzes runtime usage, and generates actionable recommendations to keep your skill ecosystem healthy.
Agent Skill Governance is the practice of maintaining, auditing, and optimizing the collection of skills an AI agent has installed. As agents accumulate dozens or hundreds of skills, governance ensures they remain organized, efficient, and free of redundancy.
Not necessarily. If you use Hermes Agent Cloud, Skill Inspector is pre-installed and runs automatically. For self-hosted setups, install via pip and run locally.
Skill Inspector is designed specifically for the Hermes Agent ecosystem. It understands Hermes skill structures (SKILL.md, references/, templates/, scripts/, assets/) and provides insights tailored to how Hermes agents organize and use skills.
Yes. Skill Inspector is Apache License 2.0 and available on GitHub. Install it with pip or use it free on Hermes Agent Cloud.
Start using Skill Inspector today — either through Hermes Agent Cloud (zero setup) or install it yourself.