Published: 2026-06-07

Hermes V0.16 Skills Hub: Browse, Scan, and Install Agent Skills Safely

Chapters / key moments (click to jump — plays here on the page)

Detailed walkthrough of the new Hermes V0.16 Skills Hub Browser — a unified interface to discover agent skills across multiple repositories (skills.sh, Claw Hub, Claw Marketplace, GitHub), preview each skill's instructions before installing, run an automated security scan for a trust verdict, and manage installed skills by tier (trusted, built-in, or community).

Source video

"New Hermes Agent Update Is INSANE" by Julian Goldie SEO — Watch on YouTube →

Key Takeaways

  • Update first: go to the Manage tab in the Hermes dashboard → scroll to Update section and install the latest version before the Skills Hub becomes available
  • The Skills Hub searches four sources at once: skills.sh, Claw Hub (OpenClaw's marketplace), Claw Marketplace, and GitHub — each result card shows a description, trust tier, and a scan button
  • Before installing any community skill, click the card to open and read the skill.md file — this reveals exactly what instructions the skill will inject into your agent's context, including any permissions it requests
  • The security scan provides an automated verdict ("safe to install" / "be cautious" / blocked) but is not a guarantee — treat it as a first filter; if a skill requests computer-use or file access, always verify the source yourself
  • Keep your installed skill list lean: too many skills increases token consumption and can confuse the agent; use the Skills tab to search your installed list and disable any you are not actively using
Security note

Community skills run inside your Hermes agent context. Always read the skill.md file before installing — look for any instructions that grant file system access, network requests, or computer use. The automated scan is helpful but not exhaustive.

What you can actually set up from this

Extracted from the video's own transcript — the specifics the original summary left out.

Commands & Code Shown

hermes update

Purpose: Update first — the Skills Hub browser needs a current Hermes (also available from the dashboard's Manage → Update).

When to use: Before using the hub.

hermes skills search <keyword>

Purpose: Search the hub by keyword (CLI equivalent of the dashboard search, per the Hermes docs).

When to use: Finding a skill for a job, e.g. email.

hermes skills inspect <source/path>

Purpose: Preview a skill before installing it.

When to use: Always, before installing a community skill.

hermes skills install <source/path>

Purpose: Install a skill; a security scan runs first.

When to use: After inspecting it.

hermes skills audit

Purpose: Re-scan all installed hub skills for security issues.

When to use: Periodically, and after updates.

Reproducible steps

  1. Search across hubs

    The hub searches skills.sh, ClawHub and other registries (and GitHub) from one box in the dashboard's Skills section.

  2. Read the verdict and the trust tier

    Each result shows a scan verdict and a colour-coded tier (built-in, trusted, community). A risky skill's install can be blocked outright.

  3. Read SKILL.md yourself

    Open the skill's SKILL.md to see exactly what instructions it gives your agent before installing.

Gotchas

  • Do not rely on the scan alone — skills can drive your browser or computer; read them.
  • More skills is not better: a large skill list slows the agent, confuses it and costs tokens. Switch off the ones you do not use.

Weekly Digest — In Your Inbox

Get the week's top AI agent news, updates, and guides — every Friday.