Skip to content
Get packs

Proof Video Recorder

Record, caption and attach real browser proof videos and screenshots to a GitHub pull request.

Source Download skill
Repository
openclaw/clawhub
Stars
9,546
Licence
MIT
Files
1 file

What it does

This skill teaches an AI agent how to produce evidence that a ClawHub web change works: it drives a real browser against a local ClawHub instance, records the session, adds captions and zooms with ffmpeg, and attaches before/after clips to a GitHub PR as native attachments. It also insists on honest labelling, such as marking failed baselines as expected failures and stating which checks were real, simulated or untested.

Great for

  • Maintainers of ClawHub or similar projects who need visual proof attached to pull requests
  • Developers who want an agent to produce reproducible before/after UI recordings
  • Reviewers who need clearly labelled evidence of a feature working end to end

Skip it if

  • Simple code changes with no visible UI, where a video adds little value and the workflow is heavy
  • Anyone without a local ClawHub setup, Playwright, ffmpeg and gh CLI, since the steps assume all of them
  • Uploading sensitive recordings to a public PR, because GitHub attachments cannot be deleted once uploaded and the skill warns to sanitize first

Try asking

  • “Record a feature proof for my ClawHub PR showing the new install button working, with captions”
  • “Make a before/after proof video of the search fix and attach it to PR 123 in openclaw/clawhub”
  • “Capture desktop and mobile screenshots of the new settings page for review”
  • “Trim idle waits in this browser recording, add a 2x zoom on the result, and burn in captions”
Step-by-step guide: how to use Proof Video RecorderInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    proof-video.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip proof-video.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip proof-video.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Name the behaviour, fixture and expected visible result before recording so the before/after comparison is fair.
  • Keep captures and browser state in ignored local folders and upload only the specific inspected files.
  • Check the full-size decisive frames yourself before attaching anything, since uploads cannot be removed.

What's inside

  • SKILL.md

Redistributed unmodified from openclaw/clawhub under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.