Downloads
browser_download {action: "list"} reports downloads: in extension mode the browser’s own download list (filtered to the agent’s tabs unless Share everything is on), in developer mode the files in the context’s download directory (~/.browspark/downloads/<context> unless overridden at launch).
{action: "wait", urlContains, timeoutMs} blocks until a matching download completes and returns its path, which is what you want right after clicking an “Export CSV” button.
browser_pdf prints the current page like Print → Save as PDF: format (A4, Letter, Legal), landscape, printBackground, scale, and path to write somewhere other than the artifacts directory. Returns the file path.
Domain policies
browser_policy restricts where a tab may navigate and load resources from. allow is an allow-list (everything else is blocked); block is a deny-list. Subdomains match, non-HTTP schemes always pass, and blocked navigations fail with ERR_BLOCKED_BY_CLIENT.
default: true the policy applies to every tab the agent opens from now on, which is the way to keep a scraping run on the sites you intend:
status shows the active policy; clear removes it. Policies are enforced through Chrome’s Fetch domain so they cover subresources, not just top-level navigation.