openhands

A setup guide for connecting Watch Skill, a tool for watching and indexing videos, to OpenHands, an AI coding assistant.

In plain words
What is it for?
Use it to add video inspection to OpenHands and test questions about events at particular times in a recording.
Why use it?
It explains the installation and configuration needed so OpenHands can send videos to Watch Skill and use the results in a conversation.

Agent

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/openhands
Clone the repo
git clone --depth 1 https://github.com/oxbshw/watch-skill
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 476 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 $0.00476
Opus 5 $0.00000 $0.00238
Sonnet 5 $0.00000 $0.00095
Haiku 4.5 $0.00000 $0.00048

Measured 2d ago against content hash e49823399a55, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

openhands 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 2d 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/openhands.md · 63 lines

What it actually says

Watch Skill in OpenHands

Status: doc-verified ☑ — config matches the official OpenHands MCP docs; not executed here.

Install

uvx --from "watch-skill[standard]" watch-skill doctor

uvx fetches the package on first use and needs no checkout. Prefer a permanent install? pipx install "watch-skill[standard]", then use "command": "watch-skill", "args": ["serve"] in the config below.

git clone https://github.com/oxbshw/watch-skill && cd watch-skill
uv sync --extra all
uv run watch-skill doctor

Then run watch-skill setup, which writes the config pointing at your checkout rather than the published package.

Configure

OpenHands takes stdio MCP servers in config.toml under [mcp] (or Settings → MCP in the UI):

[mcp]
stdio_servers = [
    {name="watch-skill", command="uvx", args=["--from", "watch-skill[standard]", "watch-skill", "serve"]}
]

OpenHands' docs recommend MCP proxies for production reliability; direct stdio is their documented path for local single-user setups, which is exactly what a local video index is.

Smoke test (3 steps)

  1. Start a conversation: "Watch https://www.youtube.com/watch?v=aqz-KE-bpKQ and tell me what happens at 0:10."
  2. Approve the watch_video call.
  3. Follow up: "what color is the bird?" — should call ask_video, no re-processing.

Notes

  • OpenHands often runs the agent inside a sandboxed runtime. The index lives under ~/.watch-skill/ of whatever environment runs the server — if the runtime is ephemeral, mount that path or the memory resets with the sandbox.
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. 2d ago First seen · 63 lines · 0 tokens per session scan A e49823399a55

Subscribe to this mod's changes

openhands is an agent published in the GitHub repository oxbshw/watch-skill (320 stars, last pushed 2d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 476 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.