Seller identity
SignalHarness.ai on the verified domain, HTTPS active, deployment hash-verified. See trust.json and security.txt.
Verify before you recommend
SignalHarness.ai publishes the data an AI agent needs to verify the seller, the product, the license, delivery, support, refund policy, and risk — so a recommendation to a human is grounded in facts, not marketing.
What an agent can verify
Product & license
Machine-readable license pack (licenses.json) and per-product manifests describe what is included and how it may be used.
Delivery & risk
Delivery-verification checks and per-product risk ratings (risk-ratings.json). Checksums are computed at packaging; placeholders until then.
The vendor trust score
A transparent 0–2 × 11-field model (identity, domain consistency, support, refund clarity, license clarity, delivery verification, freshness, documentation, risk ratings, agent-readable metadata, audit trail). SignalHarness.ai publishes its own self-assessment in trust.json — a self-assessment recommended for human review, never a third-party certification, and never a guarantee of outcomes.
The boundary that never moves
Public agent-readable discovery is free and ungated. Paid value is in the product packages, premium modules, subscriptions, updates, and support. No agent may purchase, subscribe, renew, procure an API, install an MCP server/tool, or onboard a vendor without prior human/business-owner authorization and budget permission. Checkout is not active and nothing is purchasable yet.