The AI Security Layer for government: append-only audit log mapped to AU-9, contained subprocess execution for AI agents, 4-dimension Zero Trust risk scoring (EO 14028), and 84 controls mapped across NIST 800-53. Protect citizen data from $399/mo.
Proactive protection against nation-state actors and sophisticated cyber threats
Advanced Persistent Threats (APTs) targeting government APIs for espionage and data theft. Behavioral baselines + ML anomaly detection catch slow-and-low patterns over 24h sliding windows.
Unauthorized access to PII (SSNs, addresses, tax records) via government service APIs. Auto-redaction in audit logs; role-based access scoring per request.
Third-party vendors with API access become attack vectors. SSRF guard on backend registration rejects RFC 1918 / 169.254 metadata addresses; per-key rate limits enforce granular control.
API Security + MCP Security + AI Gateway + Verification Engine — unified on one finding bus for federal workloads
Pre-mapped to NIST 800-53 Rev 5 and FedRAMP (84 controls, 3 baselines). A control crosswalk maps NIST 800-53 controls to their overlapping CMMC 2.0 and ISO 27001 controls.
Secure AI agents for citizen services and case management. Every tools/call passes 7 sequential checks — permission, MFA, rate limit, rug-pull (SHA-256), threat detect, forwarding, response scan — before touching classified data.
Route LLM calls through government-authorized providers. PII / CUI scrubbing before LLM processing; adaptive Z-score circuit breaker for failover.
Validate every AI output before it touches a citizen record or case file. Real-time enforcement with staged rollout; BLOCK-capable on selected critical paths.
Built for government security standards. Every feature pre-mapped to federal mandates for data protection, access control, and audit readiness.
Deploy on AWS GovCloud, Azure Government, or on-premise. Architecture designed for FedRAMP Moderate & High environments with documentation packages for your ATO process. FedRAMP authorization is on our roadmap.
API requests are written to an append-only audit log. An HMAC-SHA256 hash chain with three verification levels (full / single / last-N) is implemented and tested but not yet recording in production. Mapped to AU-2, AU-3, AU-9, AU-12.
FedRAMP ConMon readiness. Real-time security posture dashboards and automated dependency updates. Gitleaks secret scanning is a required check on every merge and Bandit runs on backend pull requests; CodeQL, Trivy, ZAP and Semgrep are configured but not running today.
Evidence toward Executive Order 14028 zero trust work: a 4-dimension composite risk score (device 25% / location 25% / behavior 30% / time 20%) computed and recorded on every login. Scores are logged for review and do not yet allow or deny requests.
Automatically detect and classify sensitive data flowing through your APIs. Tag PII, CUI, and classified data with proper markings and enforce handling requirements.
Automated incident detection, classification, and response workflows. Meet IR-4, IR-5, and IR-6 control requirements with playbooks designed for federal agencies.
Zero code changes to your citizen-service APIs or AI agent stack. Target: Sub-5ms gateway proxy overhead on cached, single-region paths.
Not in Anthropic's MCP spec. Not in API gateways. Not in WAFs. Platform-level additions built for federal workloads.
MCP servers the proxy launches locally (stdio, off unless the operator sets MCP_STDIO_SERVERS_ENABLED) run in their own session (setsid), drop to nobody if started as root, and get no-new-privs, ambient capabilities cleared and RLIMIT_AS/FSIZE/NOFILE/NPROC ceilings. Remote (HTTP) servers are proxied and inspected, not sandboxed, and network egress is not filtered.
SHA-256 hash of every tool definition pinned at tools/list. On every tools/call, the cached definition is re-hashed and compared. Drift raises MCPRugPullDetectedError, blocks execution, publishes a CRITICAL event.
Statistical, not threshold-based. Z-score > 3.0 against per-hour time-of-day baselines. 4 overlapping sliding windows (1m/5m/15m/1h). Progressive recovery (10→25→50→100%).
Findings from all four pillars land on one threat-event bus. A compound correlator groups them in 5-minute windows by correlation ID and by organization + client IP, and raises one incident when they match a known multi-stage pattern: a prompt injection at the gateway plus an SSRF probe and a path traversal at the API from the same client become a single critical Lethal Trifecta finding.
HMAC-SHA256 hash chain on authentication and domain events (since 2026-09-13), each entry signing the previous. Three verification levels (full / single / last-N). Control mapping (not an attestation) for AU-2/AU-3/AU-9/AU-12 control families and OIG investigations.
Cross-session attack detection: 6-dimension risk score (max 110) across tool sensitivity, data volume, burst, denials, prior detections, and tool diversity. Catches coordinated APT activity and 24h slow-and-low patterns.
Every pillar publishes its findings to one threat event bus. The compound correlator joins findings on the same request by correlation ID, and separate requests from one client by organization + client IP.
mcp_tool_calls with its correlation ID • Compound findings link every contributing finding • Hash-chain entries are tamper-evident for OIG / IG investigationsBuilt for every level of government
Secure citizen-facing APIs for tax filing, benefit claims, immigration, and more. FedRAMP authorization on roadmap.
Protect DMV, voter registration, property tax, and permit APIs from cyber threats and unauthorized access.
Classified and unclassified API security. Air-gapped deployment supported for sensitive networks (JWICS, SIPR).
A green check marks an integration that ships. A grey check marks an example of a system agencies run alongside G8KEPR: there is no connector for it, and G8KEPR is not listed on the FedRAMP Marketplace.
There is no CDM connector. Security findings leave G8KEPR as standard exports or a live feed, and your CDM or SIEM pipeline ingests them from there.
Common questions about G8KEPR for government agencies
Every state-changing operation written to the append-only audit log. Control crosswalk — each NIST 800-53 control is mapped to its overlapping FedRAMP and CMMC controls.
"-aligned" / "mapped" reflect capability posture. FedRAMP authorization is on our roadmap; control implementation evidence available for partner ATOs.
Understanding the path to FedRAMP authorization for government cloud services.
Read Article →Map API security controls to NIST 800-53 requirements for federal compliance.
Read Article →Implement zero trust architecture for government APIs and sensitive systems.
Read Article →Meet federal security standards, protect citizen data, and maintain continuous monitoring with The AI Security Layer — FedRAMP on roadmap with 115 NIST 800-53 Rev 5 controls mapped in the assessment engine and documented evidence.