hermes

A setup guide for using the Watch Skill video-analysis tools with Hermes Agent and similar coding-agent systems. Hermes is an agent framework that extends its abilities through instruction files and command-line tools.

In plain words
What is it for?
Use it to add video watching, question answering, and related Watch Skill workflows to Hermes or another agent that supports skills, project instructions, or HTTP tools.
Why use it?
It explains how to install and connect the tools when there is no built-in plugin store or direct integration.

Agent

Part of the watch-skill plugin — 10 skills, 1 command, 27 agents, 1 MCP server shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/oxbshw/watch-skill/hermes
Clone the repo
git clone --depth 1 https://github.com/oxbshw/watch-skill

Or install watch-skill, the plugin that ships this one along with the rest of its 10 skills, 1 command, 27 agents, 1 MCP server.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 482 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00482
Opus 5 $0.00000 $0.00241
Sonnet 5 $0.00000 $0.00096
Haiku 4.5 $0.00000 $0.00048

Measured 3d ago against content hash 3bac92154035, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

hermes scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

docs/agents/hermes.md · 55 lines

What it actually says

Watch Skill in Hermes Agent (and Hermes-style harnesses)

Status: doc-verified ☑ — matches the hermes-agent docs; not executed here.

Hermes Agent (Nous Research) and harnesses like it extend through skills — instruction files over CLI tools — rather than a plugin store. Watch Skill's skills library is exactly that shape.

Install

# the skills shell out to `watch-skill`, so it needs to be on PATH
uv tool install "watch-skill[standard]"   # or: pipx install "watch-skill[standard]"
watch-skill doctor

Configure

Two working paths, pick one:

Skills. Add the library to the harness's skills directory (Hermes bundles skills as directories with instruction files; copy or symlink ours):

skills/  →  your hermes skills dir

AGENTS.md. Hermes-style harnesses read repo instructions. Drop adapters/agents-md/AGENTS.md into the project root — it teaches any instruction-following agent the full CLI contract (watch → ask → loop) with no client support needed at all.

For a networked setup, Hermes exposes an OpenAI-compatible endpoint and can call HTTP tools; run watch-skill api and use the REST recipe in frameworks.md (OpenAPI at /openapi.json).

Smoke test (3 steps)

  1. Ask: "Watch https://www.youtube.com/watch?v=aqz-KE-bpKQ — what happens at 0:10?" — the agent should run watch-skill watch.
  2. Check for t=MM:SS citations in the answer.
  3. Follow up about the same video — expect watch-skill ask, not a re-watch.

Notes

  • This page doubles as the recipe for any agent whose only extension points are "instructions + a shell": skills dir if it has one, AGENTS.md if it doesn't, REST if it can't even shell out.
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 55 lines · 0 tokens per session scan A 3bac92154035

Subscribe to this mod's changes

hermes is an agent published in the GitHub repository oxbshw/watch-skill (320 stars, last pushed 3d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 482 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.