| name | browser-content-review |
| description | Read, extract, or summarize web-page content when the user explicitly asks to "summarize this page", "summarize the latest message", "extract the table", or "review this web content". |
| category | automation |
| version | 1.0.0 |
| co-load-with | ["browser-use"] |
| requires-primary | ["browser-use"] |
| co-load-triggers | ["summarize this page","summarize the latest message","extract the table","review this web content","summarize the page","read this web page"] |
| examples | [{"prompt":"Open the dashboard and summarize the latest update."},{"prompt":"Read this web page and extract the table into bullets."}] |
| test_commands | ["python -m pytest tests/test_skills.py tests/test_prompts.py tests/test_mcp_client.py -q"] |
Browser Content Review
Scope
Use alongside Browser Use when the user asks to read, extract, compare, or
summarize content already visible in a web page or web app. Do not use it for a
generic browser navigation request.
Evidence-First Reading
- Navigate and take a Playwright snapshot; use the current page text or a
targeted read-back as the source for the summary.
- Capture the requested facts, dates, counts, names, and current page state.
Preserve uncertainty when text is incomplete or the page has not loaded.
- Summarize only what was observed. Do not infer hidden messages, unread
content, delivery status, or information outside the current view.
- Keep the answer short and structured around the user's question. Quote only
the minimum necessary text; redact private contact values.
Boundaries
- Reading/summarizing does not authorize replying, liking, posting, sending, or
changing the page. Those require clear user intent and their own verification.
- If the page changed or a target is stale, obtain one fresh snapshot before
continuing rather than guessing from prior references.