ClawHub is a public registry where OpenClaw users publish, version, search, and install text-based agent skills and OpenClaw packages. It provides web browsing, a CLI-oriented API, moderation, vector search, and artifact hosting for code plugins, bundle plugins, and experimental whole-agent packages. The catalogue skills and agents are entries that can be discovered or used through this registry.
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.
npx skills add openclaw/clawhub --skill proof-videogit clone --depth 1 https://github.com/openclaw/clawhubWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/openclaw/clawhub/proof-video)<a href="https://agentmods.dev/skills/openclaw/clawhub/proof-video"><img src="https://agentmods.dev/badge/skills/openclaw/clawhub/proof-video/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/openclaw/clawhub/proof-video"><img src="https://agentmods.dev/badge/skills/openclaw/clawhub/proof-video.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00039 | $0.01651 |
| Opus 5.5 | $0.00016 | $0.00660 |
| Sonnet 5.5 | $0.00008 | $0.00330 |
| Haiku 4.5 | $0.00004 | $0.00165 |
Grade A, and why
proof-video 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 16d 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.
How it starts
The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Proof video
Adapted from OpenClaw's proof-video skill and PR evidence instructions. This adaptation uses ClawHub's existing real-browser proof runner; OpenClaw's mocked Gateway template and repo-specific helper imports do not apply here.
Establish the claim
Before recording, name the behavior, exact source refs, local URL, fixture and expected visible result. Keep the same viewport, theme, account, data and interaction sequence for before/after comparisons. Feature proof can omit an impossible baseline, with the reason stated.
Use a real ClawHub instance with the relevant local Convex code and fixture state. Use a real isolated browser on the current host; use Crabbox only when its environment is needed and available. Do not use HTML mockups, synthetic screenshots, intercepted success responses or hand-written product UI as proof. Seed through the supported local fixture workflow. Never use the operator's personal Gateway for cross-system proof.
For a ClawHub → OpenClaw flow, record the actual admin action, security pending state, catalog lookup, installation and meaningful live tool result. State separately which scan/provider checks are real, simulated or untested. A queued job is not a completed scan, and discovery is not account-level authorization.
Capture
Keep captures, scripts, cue times and private browser state in separate ignored local directories. Select individual inspected media files for upload; never upload the entire artifact directory or browser state.
Use the existing runner with a task-specific scenario:
bun run proof:ui -- --runner local --mode feature \
--scenario .artifacts/proof-scenarios/example.pw.ts \
--candidate-url http://127.0.0.1:3000
Use --mode before-after, --baseline-url and --candidate-url for paired proof. See scripts/ui-proof-runtime.mjs for the scenario contract and specs/ui-proof.md for runtime isolation. For a multi-app flow that needs a dedicated driver, use the existing Playwright dependency and recordVideo against the real services.
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.
- 16d ago First seen · 84 lines · 39 tokens per session scan A 9fc16b352ca5
proof-video is a skill published in the GitHub repository openclaw/clawhub (9,517 stars, last pushed today), licensed MIT. It adds 39 tokens to every session and 1,651 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-10-02.
Other skills, from other repositories
qa/e2e-playwright
A method for writing Playwright end-to-end tests, which automate a browser to verify complete user journeys. It covers setup, reusable page objects, login reuse, API mocking, visual checks, multiple browsers and screen sizes, CI, and debugging.
directory-based-skill
A skill loaded from a directory structure with SKILL.md frontmatter, plus file-based and URL-based instruction sources.
chatgpt-bridge
A bridge that lets an agent ask questions through a user's already signed-in ChatGPT browser session and retrieve the answers. It can also find an earlier ChatGPT conversation and continue it.
xiaohongshu-skill
A skill for working with Xiaohongshu, also called RedNote, a Chinese social-media platform for posts, videos, articles, comments, and reactions. It uses a browser session and returns structured results.
firefox-control
Use to automate web browsing via Firefox WebDriver BiDi — tab listing, JS evaluation, screenshots, DOM helpers.
chrome-control
Use when debugging web applications, diagnosing page errors, inspecting console output, or capturing screenshots via Chrome DevTools Protocol (CDP).