thkt use-workflow-pageshot
Input
PR body is passed as a string from /pr. If either field below is missing, return mode=failed reason=<missing field name> and let /pr decide.
| Field | Source |
|---|---|
| Preview URL | The leading Preview URL: <URL> line in the PR body |
| Steps | Numbered list under the ## How to Test section in PR body |
Browser engine
Pass --session pageshot --engine chrome on every agent-browser call. The default lightpanda returns a placeholder PNG for screenshot.
Mode Selection
| Steps count | Mode | Artifact |
|---|---|---|
| 1 | screenshot | step-01.png |
| 2+ | video | capture.mp4 |
Execution
| Step | Action |
|---|---|
| 1 | Create the output directory .claude/workspace/pageshot/$(date +%Y%m%d-%H%M%S)/ |
| 2 | agent-browser open {Preview URL} |
| 3 | Run Screenshot Flow or Video Flow based on Mode |
| 4 | Print absolute path of artifact to stdout |
Screenshot Flow (1 step)
| # | Command |
|---|---|
| 1 | agent-browser snapshot to obtain accessibility tree |
| 2 | If the step contains operations, run agent-browser {click/type/fill/...} |
| 3 | agent-browser screenshot --full {outdir}/step-01.png |
Video Flow (2+ steps)
Insert agent-browser wait 500 between steps. Run snapshot before each operation to identify elements.
| # | Command |
|---|---|
| 1 | agent-browser record start {outdir}/capture.webm |
| 2 | For each step run snapshot then the operation in order |
| 3 | agent-browser record stop |
| 4 | ffmpeg -i {outdir}/capture.webm -vcodec libx264 {outdir}/capture.mp4 |
Output
Return a single mode line on stdout. /pr branches on that line.
| Situation | Line |
|---|---|
| A screenshot was taken | mode=screenshot artifact=/absolute/path/to/step-01.png |
| A video was taken | mode=video artifact=/absolute/path/to/capture.mp4 |
| An input was missing | mode=failed reason=<missing field name> |
| The capture failed | mode=failed reason=<one-line reason> |