API Reference
All endpoints return JSON. Rate limited to 60 requests/minute per IP. No authentication required.
Rate Limiting: 60 requests/minute per IP. Responses include X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers. HTTP 429 when exceeded.
Search
POST/api/search
Search the trust-scored agent capability index.
curl -X POST https://agentsearch.cybersecai.co.uk/api/search \
-H "Content-Type: application/json" \
-d '{
"query": "payment processing",
"minTrust": 0.5,
"protocol": "mcp",
"maxResults": 10
}'
GET/api/search?q=payment
Quick search via query parameter.
Verify
POST/api/verify
Independently verify a source's trust status.
curl -X POST https://agentsearch.cybersecai.co.uk/api/verify \
-H "Content-Type: application/json" \
-d '{ "sourceId": "mcp-secure" }'
Warnings
GET/api/warnings?sourceId=xxx
Check warnings (CVE, signature issues) on a source. Trust score is never altered by warnings -- labels are informational only. Your agent decides.
Claim & Report
POST/api/claim
Claim ownership of a source. Submit your public key, receive a cryptographic challenge.
curl -X POST https://agentsearch.cybersecai.co.uk/api/claim \
-H "Content-Type: application/json" \
-d '{
"sourceId": "my-mcp-server",
"publicKey": "-----BEGIN PUBLIC KEY-----..."
}'
POST/api/claim/verify
Complete ownership verification by signing the challenge with your private key.
POST/api/report
Report a malicious or fraudulent source. Requires sourceId and reason.
POST/api/remove
Request removal from the index. Verified owners get instant removal.
POST/api/withdraw
Owner-initiated withdrawal. Adds OWNER WITHDRAWN label. Requires signed proof.
Stats
GET/api/stats
Index statistics: total indexed, signed count, blocked count, search count.
GET/api/crawl-stats
Crawler statistics: last crawl time, signed vs unsigned breakdown.
SDK Verification
GET/api/verify-sdk
Verify the @proofxhq/agentsearch npm package integrity. Returns ECDSA public key, SHA-512 hash, and signature.
MCP Server
GET/mcp/tools/list
MCP tool discovery. AgenticSearch is itself an MCP server. Any MCP-capable agent can search through standard tool calls.
// Tools available:
search -- search the index
verify -- verify a source
claim -- claim ownership
report -- report a source
remove -- request removal
revoke -- owner withdrawal
check_revocation -- check warnings
Health
GET/health
Health check. Returns indexed count and search count.