All nineteen MCP tools
Canonical list of the local Listaro MCP tool names and their responsibilities.
The names below match the Implementation 1 src/mcp/contracts.ts registry. The local server offers these tools; exact positive and negative certification results must still be independently reconciled from the final engine report.
Discovery
search_web— find results using a configured authorized provider. Without one, a typed provider-not-configured response may be returned.map_site— discover site URLs within bounded scope and sitemap limits.list_extraction_presets— inspect supported extraction presets.
Collect and process
scrape_page— collect content, metadata, links and source evidence for one URL.scrape_batch— batch collect URLs with explicit concurrency.crawl_site— create a bounded durable website crawl.parse_document— parse supported PDF, DOCX, XLSX, HTML and text sources.browser_session— create, inspect or close an isolated session.browser_action— perform an authorized session action.
Inspect and control
get_job— inspect job status and error/progress evidence.cancel_job— request idempotent cancellation of a queued/running job.get_crawl— page through crawl results.get_artifact— retrieve verified, bounded artifact content.
Monitor and compare
create_watch— create a saved target for future checks.check_watch— run an explicit watch check.set_watch_schedule— enable or pause recurring checks; scheduling defaults to off.list_watches— inspect watches.compare_crawls— compare existing crawl snapshots.get_change_report— retrieve a bounded change feed/report.
Contracts and compatibility
The canonical Zod inputs, default values and schema constraints live in src/mcp/contracts.ts in the repository. Do not substitute a marketing sample for that source. The 1B enhancement contract must be additive and preserve these names and existing required arguments.