Docs · hypothesis-led

Agents and LLMs

Start from /llms.txt and /.well-known discovery. Three free standalones are callable without auth (rate-limited): Design UX Radar, Search Repair teaser, Deployment Trust packet. Workspace missions need a founder token. Never treat discovery as permission to publish, deploy, send, or free-form spend.

Open agent contract Human quickstart

Docs hub Get started llms.txt Agent capabilities

Hypothesis-led documentation. No ranking, traffic, or outcome guarantees. Public discovery is not execution authority.

Read order for agents

/llms.txt (includes free-tool recipes) → /.well-known/agent-capabilities.json → /.well-known/mcp.json + a2a.json → /use-from-agent operating contract → challenge manifest only with a founder-provided token.

Public free standalones (no bearer)

MCP tools: launchfiles_standalone_design_ux_radar_free (url), launchfiles_standalone_search_repair_radar_free (gsc_site_url; owned adapter GSC bind needs no access_token; stranger sites need_auth without token), launchfiles_standalone_deployment_trust_radar_free (url, use_case_plain, autonomy, data_sensitivity). A2A methods: design_ux_radar.run_free, search_repair_radar.run_free, deployment_trust_radar.run_free. Free tier only — never unlocks paid vision/model SKUs.

Paid agent path

List providers / create checkout (discover only), poll until paid, receive lf_agent_ key once, then design_ux_radar.run_paid / launchfiles_standalone_design_ux_radar_paid with credit debit. Rails: stripe | xmoney | x402 when enabled.

Forbidden without explicit grant

Publish, deploy, send, free-form spend, billing ledger mutation outside checkout, gate bypass, unrestricted tool access, and secret exfiltration.

Next safe action

Learn the contract, then run a bounded mission.

Prefer receipts over claims. Prefer fail-closed honesty over impressive incompleteness.

Quickstart