{"schema_version":"1.0","protocol_version":"2025-03-26","name":"agentguard","display_name":"AgentGuard","version":"3.0.0","description":"OWASP LLM06 Loop Circuit Breaker & Stripe AI Margin Firewall. AgentGuard wraps your MCP servers and LLM gateways in <1ms · halting >4x/60s semantic tool-call loops with JSON-RPC -32006 circuit breaking, gating destructive tools behind Slack/Discord approval, and auto-downgrading negative-margin Stripe tenants.","homepage":"https://agentguard.audiscape.com","transport":{"type":"streamable-http","url":"https://agentguard.audiscape.com/mcp"},"endpoint":"https://agentguard.audiscape.com/mcp","documentation_url":"https://agentguard.audiscape.com/llms.txt","endpoints":{"mcp":"https://agentguard.audiscape.com/mcp","mcp_jsonrpc":"https://agentguard.audiscape.com/mcp","rest_scan":"https://agentguard.audiscape.com/api/scan","rest_checkout":"https://agentguard.audiscape.com/api/checkout","rest_bundle":"https://agentguard.audiscape.com/api/bundle"},"authentication":{"type":"none_required_for_audit","paid_bundle_header":"Authorization: Bearer ag_live_<key>"},"tools":[{"name":"audit_domain","description":"Run a live HTTP audit against a public domain and return score, grade and report URL.","inputSchema":{"type":"object","properties":{"domain":{"type":"string","description":"Target domain, e.g. cursor.com"},"mode":{"type":"string","enum":["citerank","checkoutguard","agentguard","signalstrike"]}},"required":["domain"]}},{"name":"get_remediation_preview","description":"Return the remediation preview and scorecard URL for a domain.","inputSchema":{"type":"object","properties":{"domain":{"type":"string"}},"required":["domain"]}},{"name":"create_checkout_session","description":"Return a Stripe checkout link for the remediation pack for a domain.","inputSchema":{"type":"object","properties":{"domain":{"type":"string"},"mode":{"type":"string","enum":["citerank","checkoutguard","agentguard","signalstrike"]}},"required":["domain"]}},{"name":"guard_check_session","description":"Replay an agent's tool-call log through the AgentGuard loop and budget guard and decide whether the next call should run. Catches identical and near-repeat loops (same file, URL or command), re-reads of finished items, failing-tool retry streaks and USD or token caps before spend.","inputSchema":{"type":"object","properties":{"calls":{"type":"array","maxItems":20000,"items":{"type":"object","properties":{"tool":{"type":"string"},"arguments":{},"isError":{"type":"boolean"},"costUsd":{"type":"number","minimum":0},"totalCostUsd":{"type":"number","minimum":0,"description":"Running session spend so far; spend becomes max(spend, total)"},"totalTokens":{"type":"number","minimum":0,"description":"Running session tokens so far"},"tokens":{"type":"number","minimum":0},"errorText":{"type":"string","description":"Error message when isError; permanent errors (missing binary, 4xx, quota, permission) halt after 1 retry"},"readOnly":{"type":"boolean","description":"Call had no side effects; lets the opt-in completed-read rule apply"},"resultHash":{"type":"string","description":"Stable hash of the tool result; repeats with the same hash count as no progress"}},"required":["tool"]}},"next":{"type":"object","properties":{"tool":{"type":"string"},"arguments":{}},"required":["tool"]},"config":{"type":"object","description":"Optional agentguard.json contents"}},"required":["calls","next"]}}],"errorCodes":{"-32006":"Loop: identical call repeated, or opt-in near-repeat target / completed re-read (G5, G5a, G5b)","-32007":"Destructive tool without approval","-32008":"Budget cap reached (per-call costs or running totals)","-32009":"Error streak, or permanent error retried on the same tool and target (G4, G4a)","-32010":"Session halted after repeated blocks","-32011":"Idle burn: opt-in, no new results for loop.maxNoProgressCalls calls (G4b)"}}