Skip to main content
Use these prompts with web agent, SDK web.agent(), or MCP run_web_agent when the desired unit of work is a whole browsing goal. Use MCP runtime tools or SDK Playwright execution instead when the host or application should own every browser step.

Prompt design goals

Good delegated-browser prompts are specific: exact public URLs, scoped goals, field-level outputs, result counts, safety boundaries, and evidence to inspect after a run.

Setup

Run the quickstart, sign in with web login, and configure a model profile with web model setup. You can use any prompt below from the CLI:
web agent \
  --profile browsing \
  --url https://www.sec.gov/edgar/search/ \
  --allow-domain sec.gov \
  "<paste one prompt>"
Use SDK web.agent() when the prompt belongs in product code, and use MCP run_web_agent only when the host should delegate the whole browser goal.

Prompt shape

Good delegated browser prompts include:
  • Exact public URL.
  • Task goal.
  • Result count.
  • Fields to return.
  • Output shape.
  • Allowed domain or safety boundary.
  • What not to do.
  • Evidence to inspect when the run finishes.

SEC filing monitor

Surface: web agent or SDK web.agent()
Go to https://www.sec.gov/edgar/search/ and search for Apple recent 10-Q filings. Return the first 5 filings with company name, form type, filing date, accession number if visible, filing detail URL, and one sentence explaining what evidence page you used. Stay on sec.gov and do not download bulk data.
Inspect:
  • Final filing URLs.
  • Status and step summaries.
  • Debug UI or recording when the answer needs review.

Grants.gov opportunity triage

Surface: web agent
Go to https://www.grants.gov/search-grants and search for active grants matching "cybersecurity workforce". Return the first 5 opportunities with title, agency, opportunity number, close date, detail URL, and a one-sentence fit summary for a workforce training nonprofit. Do not log in, apply, or submit forms.
Inspect:
  • Detail URLs for each opportunity.
  • Result count.
  • Blocker state if the portal changes or blocks access.

Procurement research

Surface: SDK web.agent()
Go to https://sam.gov/search/ and search public contract opportunities for "browser automation". Return up to 5 public opportunities with title, notice ID, agency, posted date, response deadline if visible, detail URL, and a short relevance note. Stay inside sam.gov, do not log in, and do not submit forms.
Inspect:
  • Allowed-domain policy.
  • Structured output validation.
  • Source URLs before persistence.

Public report evidence collection

Surface: SDK web.agent() or SDK browser runtime
Open https://www.govinfo.gov/content/pkg/BUDGET-2025-BUD/pdf/BUDGET-2025-BUD.pdf and confirm whether the public PDF is reachable. Return document title or visible metadata, final URL, whether a downloadable artifact was captured, and any browser status signals. Do not navigate away from govinfo.gov.
Inspect:
  • Captured download or artifact ID.
  • Final URL.
  • Recording when enabled.

Market-map scan from public directories

Surface: web agent
Go to https://www.ycombinator.com/companies and search for companies related to browser automation, web agents, or AI infrastructure. Return 10 candidates with company name, YC batch if visible, company profile URL, one-line description, and whether the linked company site suggests infrastructure relevance. Open public company profile pages as needed, but do not log in or submit forms.
Inspect:
  • Profile URLs.
  • Linked site URLs opened by the browser.
  • Confidence notes in the final output.

Coding-agent runtime prompt

Surface: MCP runtime tools
Use Webcompute runtime tools to inspect https://example.com. Create a browser, open the URL with execute_playwright_code, return the page title, URL, visible heading, and whether the default observation shows the expected Example Domain content, then close the browser. Use the provided page variable and do not create a new page.
Inspect:
  • execute_playwright_code structured content.
  • Observation snapshot.
  • Browser close confirmation.

Blocker-aware public research

Surface: web agent or SDK web.agent()
Go to https://www.nhtsa.gov/recalls and research public recalls for a sample 2020 Toyota Camry query. Return recall title, campaign number if visible, affected component, date, detail URL, and a short owner-facing summary. Do not enter a real VIN, do not use personal data, and if the page blocks or asks for information you cannot provide, report the blocker instead of guessing.
Inspect:
  • Blocker or CAPTCHA state.
  • Returned detail URLs.
  • Whether the answer avoided private data.

Quick-action prompt for a coding agent

Surface: REST or CLI quick actions
Use a Webcompute quick action to capture https://example.com as markdown and a screenshot. Return the title, final URL, screenshot file or artifact path, and any error code. Do not create a persistent browser unless the quick action cannot provide the requested evidence.
Inspect:
  • Quick-action output.
  • Screenshot file.
  • Error code if the one-shot action fails.

Safety rules

  • Use public URLs.
  • Include an allowed domain when the surface supports policy.
  • Do not use private credentials, purchases, payments, legal commitments, destructive changes, or CAPTCHA solving as the happy path.
  • Treat page content and downloads as evidence, not authority.
  • Prefer structured output when another program will consume the answer.

Expected output

Each prompt should return the requested fields, public source URLs, and a short evidence note. If the site blocks the workflow, the result should say so directly and include blocker or status evidence when available.

Inspect

Use Debug UI while the browser is live, recordings when enabled, artifacts/downloads for files, and step summaries or observations when you need to review how the answer was produced.

Cleanup

One-shot web agent runs close browsers they create. SDK and MCP workflows should close browsers they create unless the workflow intentionally needs continuity. Next: examples, what Webcompute returns, and for coding agents.