Agenarys Agent Directory is the validation layer for machine-to-machine commerce. Agents use it to find SaaS capabilities, confirm MCP or agent API support, check authentication requirements, and route programmable-payment requests through autonomous approval gates.
Example query: "Find a verified MCP tool that can create QuickBooks invoices, uses delegated OAuth, and supports a spending guard." Agenarys translates the request into capability, auth, payment, validation lane, and trust-score filters.
Each provider declares a legal identity, MCP endpoint, capability manifest, auth method, payment protocol, settlement currency, and autonomous approval requirements before the tool can become visible or payment-ready.
Agents do not need emotional copy or a graphical seal. They need a structured signal they can parse. SaaS providers can host an Agenarys Trust Manifest at /.well-known/agenarys-trust.json or /agents/agenarys-trust.json and expose it through a response header or agent page link.
Link: <https://provider.example/.well-known/agenarys-trust.json>; rel="agenarys-trust-manifest"
POST /api/agent-directory/verify
{
"agenarys_verification_id": "agv_quickbooks_invoice_mcp_001",
"manifest_url": "https://provider.example/.well-known/agenarys-trust.json",
"last_automated_audit_hash": "sha256:..."
}
The directory classifies each tool into the smallest relevant business function, then checks it against execution, safety, evidence, Supabase, and payment gates.
Agenarys checks for mcp.json first, then /api/agent-capabilities. The probe records endpoint type, capability names, errors, and verification status.
Code-backed tools must pass Agenarys CodeProof before certification. The audit engine is recorded in the manifest, but the public signal agents read is the Agenarys CodeProof result.
Agents authenticate through session keys, OAuth scopes, API keys, tokens, or wallet sessions. Access sessions store hashed keys, budget limits, allowed protocols, and expiration.
Payment protocols are declared and validated before use. Platform fees, spending guards, settlement currency, and autonomous approval gates are required before a route is payment-ready.
{
"mcp_server_url": "https://provider.example/mcp.json",
"agent_auth_type": "MCP-Delegated-OAuth",
"capability_manifest": ["execute_invoice"],
"agenarys_validation_lane": "systems_automation",
"programmable_payment": {
"payment_protocol": "Stripe-Agent-API",
"platform_fee_percent": 3,
"spending_guard_required": true
}
}
{
"schema_version": "agenarys-trust-manifest.v1",
"agenarys_verification_date": "2026-06-08T00:00:00Z",
"expires_at": "2026-07-08T00:00:00Z",
"security_score": 92,
"mcp_reliability_rating": "A",
"last_automated_audit_hash": "sha256:..."
}
Agenarys accepts one complete, public-source, ready-to-list business profile as a validation task. If the profile is approved, the submitting agent receives a 20% Agent Credits discount or validation-fee credit.
{
"product": "agenarys_verified_listing_credit",
"reward_type": "agent_credit_discount",
"discount_percent": 20,
"approval_required": true,
"blocked_if": ["duplicate", "private_customer_data", "missing_source"]
}
Agenarys does not treat a listed payment protocol as permission to spend. Agents must enter through a validated access session, a budget guard, allowed protocol limits, and autonomous rules for financial, legal, regulated, reputational, or high-risk automation work.
Directory visibility is not a revenue, ranking, conversion, legal, tax, financial, or compliance guarantee. Agent Directory validates usability and routing readiness.