Security at Statio
This page is written to be read during a security review. It covers what happens to a credential from the moment you vault it, where policy is evaluated, how tenant boundaries are enforced, and what Statio has not yet been assessed on.
Credential encryption
All API keys stored in Statio are encrypted at rest with AES-256. Keys are decrypted only at call time within the gateway process and never cached or logged in plaintext. At no point does an agent see the raw credential, Statio injects it on the agent's behalf.Token expiry
STATIO_TOKENs expire after 24 hours and auto-refresh through the SDK. If a token is compromised, it's useless within a day. Revocation is immediate: kill a token from the dashboard and all in-flight requests with that token are rejected.Access policies
Access is granted per tool, not per server: an agent allowed to read charges is not thereby allowed to issue refunds. Policies are evaluated at the gateway before any credential is decrypted, and the default is deny, so an agent reaches only the tools you named.Audit trails
Every call through Statio is logged with the identity behind it, a timestamp, the tool and endpoint, the request parameters, the latency and the outcome. Logs are immutable and append-only. Exportable for compliance reviews, incident response, and forensics.Multi-tenancy
Each customer's credentials are fully isolated. Tenant boundaries are enforced at the gateway layer. There is no code path where one tenant's request can access another tenant's credentials. Isolation is architectural, not just logical.Network security
All traffic to and from Statio is encrypted with TLS 1.3. The gateway runs on GCP Cloud Run with automatic scaling, built-in DDoS protection, and no persistent server state. Requests are processed in ephemeral containers that are destroyed after execution.Vulnerability scanning
Every MCP server in the Statio catalog is scanned for known vulnerabilities using Grype before deployment. Servers with critical or high-severity CVEs are blocked from production. Scans run on every build, not only when a server is first published.Static analysis
Semgrep runs custom security rules against every MCP server before deployment. Detects sensitive environment variable reads, outbound calls to non-allowlisted domains, dynamic code execution (eval, Function, child_process), and data exfiltration patterns like Base64-encoded env vars.Network egress controls
MCP servers run on Cloud Run with VPC egress restrictions. All outbound traffic routes through a VPC with firewall rules and Cloud NAT. Servers cannot reach arbitrary domains, only allowlisted API endpoints are reachable.Binary authorization
Container images must be signed by our build pipeline before Cloud Run will execute them. No unsigned or tampered images reach production. Enforced at the GCP infrastructure level, not the application level.Compliance
Statio has no completed compliance certification yet, and we will not imply one. What exists today: full audit logs exportable in standard formats for compliance review, and an architecture: encrypted credentials, immutable append-only audit trails, resource-based RBAC, architectural tenant isolation, built to meet those requirements when we are assessed. If a certification is a hard requirement for you, tell us and we will say honestly where we are.Responsible disclosure
If you discover a security vulnerability, please report it to dans@romans.dev. We take every report seriously and will respond within 48 hours. We do not pursue legal action against good-faith security researchers.Connect your first server.
Three MCP servers and 5,000 calls a month, free and without a card. If you outgrow it the next plan is $19, and it stays $19 as the team grows.- Keys are encrypted at rest and never handed to an agent
- Cancel anytime. You drop to Free, nothing is deleted
- No card required, and no way to spend money on Free