<< All versions
Skill v1.0.0
currentAutomated scan100/100patleeman/task-factory/capture-screenshot
──Details
PublishedNovember 4, 2026 at 05:13 AM
Content Hashsha256:b948d94d35de3ebe...
Git SHAb892deab6cc9
──Files
Files (1 file, 1.5 KB)
SKILL.md1.5 KBactive
SKILL.md · 54 lines · 1.5 KB
name: capture-screenshot description: Capture a browser screenshot for task validation and attach it to the active task using attach_task_file. allowed-tools: Bash(agent-browser:*), attach_task_file metadata: author: task-factory version: "1.0" type: follow-up hooks: pre-planning,pre,post
Capture Screenshot
Capture visual proof of the current app state and attach it to the task.
Workflow
- Open the app
- Prefer
http://localhost:3000unless task context specifies another URL. - If needed, start the app first and wait until it is reachable.
- Take screenshot to a known path
bash
mkdir -p .pi/tmp/screenshotsagent-browser open http://localhost:3000agent-browser wait --load networkidleagent-browser screenshot .pi/tmp/screenshots/validation-$(date +%s).png
- Attach screenshot to task
- Use
attach_task_filewith the screenshot path. - Include
taskIdif you can read it from the task context; otherwise omit it.
Example tool call:
text
attach_task_file({path: ".pi/tmp/screenshots/validation-12345.png",taskId: "<TASK_ID>",filename: "validation-screenshot.png"})
- Report result
- Confirm which URL was captured.
- Confirm the attachment filename/ID returned by
attach_task_file.
Notes
- Use a deterministic screenshot path (don’t rely on unknown temp output paths).
- If capture or attach fails, explain exactly what failed and stop.
- You may take multiple screenshots if needed, but attach at least one clear validation image.