agentleFS
everrunseverruns/everruns/.agents/skills/ui-screenshots/SKILL.md

ui-screenshots

Skill67 starsChanged 16 days agoMITView on GitHub

Take UI screenshots using agent-browser. Use this skill to capture visual state of UI components for code review, visual regression testing, or documentation.

  • Reads credentials
  • Installs packages

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 0 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Blank screenshot: wait for networkidle before capturing.

    line 62
  2. Very specific:

    Use AGENT_BROWSER_ARGS rather than --args: --args splits on commas, so a multi-value flag like --disable-features=A,B becomes two argv …

    line 77

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

UI Screenshots

Capture UI state with agent-browser. Review evidence is never committed: the helper uploads it to GitHub's user-attachments CDN and embeds it in a PR comment. Curated product assets are the exception. The maintained demo set is committed under assets/screenshots/ and follows the scene contract in knowledge/ui/demo-screenshots.md. The general media and video rules live in ../ship/references/pr-and-merge.md.

Scripts

Each script documents its own usage and requirements in its header; read it if the flags matter.

.agents/skills/ui-screenshots/scripts/check-config.sh                       # GitHub auth, agent-browser
.agents/skills/ui-screenshots/scripts/take-screenshot.sh <URL> <OUTPUT>
.agents/skills/ui-screenshots/scripts/capture-demo-screenshots.sh [URL] [OUTPUT_DIR] # 2x HiDPI
.agents/skills/ui-screenshots/scripts/upload-screenshot.sh <PATH> <PR> [DESCRIPTION]

For anything the scripts do not cover, drive agent-browser directly (open, screenshot --full, snapshot -i -c, scroll, --session <name> to isolate instances).

Setup

npm install -g agent-browser
agent-browser install            # add --with-deps on Linux when system libs are missing

The maintained demo capture uses the repository's Playwright installation so it can set DPR 2 without changing the 1440 by 900 CSS viewport. Install apps/ui dependencies and its Chromium build before running it:

cd apps/ui
pnpm install --frozen-lockfile
pnpm exec playwright install chromium

Uploading needs GITHUB_TOKEN or an authenticated gh CLI session. The token must have push access to everruns/everruns.

Non-obvious failures

  • Missing browser build (e.g. chromium_headless_shell-1208) with storage.googleapis.com unreachable: symlink a nearby version in /root/.cache/ms-playwright/ — minor version drift (1200 vs 1208) is normally compatible.
    cd /root/.cache/ms-playwright && ln -s chromium_headless_shell-1200 chromium_headless_shell-1208 && ln -s chromium-1200 chromium-1208
  • Page hangs on localhost: the dev server is not up. See the local dev commands in AGENTS.md.
  • Blank screenshot: wait for networkidle before capturing.
  • Every HTTPS page fails with ERR_CONNECTION_RESET from a cloud agent, while curl through the same proxy works: Chrome's TLS 1.3 ClientHello carries a post-quantum key share that pushes it to ~2 KB, and the egress relay accepts the CONNECT then cuts the tunnel mid-handshake. curl and openssl s_client never offer a PQ key share, which is why the proxy tests healthy by hand. take-screenshot.sh already passes --ssl-version-max=tls1.2; when driving agent-browser yourself, export the flags instead of passing them per command (EVE-807):
    export AGENT_BROWSER_ARGS=$'--ssl-version-max=tls1.2\n--disable-features=PostQuantumKyber'
    Confirm they actually reached Chrome — a daemon relaunch silently drops flags given only on the first open, and the symptom is indistinguishable from the bug itself:
    pgrep -a chrome | tr ' ' '\n' | grep ssl-version
    Use AGENT_BROWSER_ARGS rather than --args: --args splits on commas, so a multi-value flag like --disable-features=A,B becomes two argv entries and Chrome exits with Multiple targets are not supported in headless mode.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.