
Pre-Inference MCP & Agent Firewall
Inspects every Model Context Protocol (MCP) tool invocation, enforcing per-environment allowlists and blocking indirect prompt-injection payloads before execution.
Protect production and govern the AI agents you ship. (Volumi personalizzati, residenza dati UE/UK, SSO/SCIM e gestione multi-tenant delegata.)

Inventory every agent, model, owner and MCP server in one place.
Define exactly which tools each agent may call, per environment.
Detect indirect injection in user input and retrieved content.
API keys, tokens and personal data are removed before inference.
Writes, payments and shell commands are routed for step-up approval.
Every allow, block and approval is recorded with its rationale.

Inspects every Model Context Protocol (MCP) tool invocation, enforcing per-environment allowlists and blocking indirect prompt-injection payloads before execution.

Automatically redacts bearer tokens, API keys and personal data before model inference and writes every allow/block decision to an append-only ledger.
CSH-RULE-001..004) mapped to MITRE ATT&CK (T1110.001, T1098.003, T1048, T1190) with idempotent dedupe_key ingestion and SHA-256 immutable event hashes.[REDACTED_SECRET], [REDACTED_PII]), explicit per-agent Model Context Protocol (MCP) tool allowlists, and indirect prompt-injection blocking.ENABLE / FORCE ROW LEVEL SECURITY across all 20 tenant tables bound to transaction-local current_tenant_id(); cross-tenant queries fail closed with HTTP 403.isolate_host) and credential revocation require named human step-up approval recorded in an append-only audit trail.DNS TXT or HTTP ownership token required before scanning. Continuous evidence aligned with Agenzia per la Cybersicurezza Nazionale (ACN), Direttiva NIS2 Italia, GDPR.ML-KEM / ML-DSA), and 2035 full PQC completion.FORCE ROW LEVEL SECURITY on all 20 tenant tables.ALTER TABLE security_events ENABLE ROW LEVEL SECURITY; ALTER TABLE security_events FORCE ROW LEVEL SECURITY; CREATE POLICY tenant_isolation_policy ON security_events FOR ALL TO csh_app_role USING (organization_id = current_tenant_id()) WITH CHECK (organization_id = current_tenant_id());
No. Il rilevamento usa regole deterministiche. ARYA SAI produce solo riepiloghi consultivi e ogni contenimento richiede approvazione umana.
Scegliete tra UE (Francoforte) o Regno Unito (Londra) in fase di attivazione.
Sì, CSH Compliance collega automaticamente le prove tecniche ai controlli ISO/IEC 27001:2022, NIS2 e GDPR.
Solo gli asset di cui avete verificato la proprietà tramite DNS TXT o token HTTP.
Sì, con tenant delegati e licenze condivise.

30 minuti con un security engineer. Niente slide.