AI Agent Evaluation

AI Security & Agent Validation

Validate AI agents, LLM-powered applications, MCP servers, AI endpoints, memory, context, and runtime audit trails.

7Suites
0Plugins
0Enabled
Plugin detection active
Built-in fallback probes active
Runtime adapters safe by default

Validation Suites

Suite 21 - AI Agent Security

Validates owned AI agents against prompt injection, jailbreaks, unsafe tool use, function-calling abuse, hallucination risk, and multi-agent trust boundary issues.

Plugin ready
Prompt Injection ResistanceIndirect Prompt InjectionJailbreak ResistanceTool Misuse DetectionTool PoisoningUnsafe Function CallingAgent Permission EscalationHallucination Risk AssessmentMulti-Agent Trust Boundaries

Plugin hooks

AgentGG

Optional deep AI code and agent behavior analysis for Suite 21.

Sandyaa

Authorized validation workflows for owned AI agents in Suite 21.

Custom AI Rules

Repository-local rules that extend AI Agent Evaluation without changing platform code.

Suite 22 - AI Infrastructure & Endpoint Security

Discovers and validates exposed AI servers, MCP endpoints, gateways, inference APIs, authentication, rate limiting, and AI service fingerprinting.

Plugin ready
Exposed Ollama ServersMCP Server ExposureAI Gateway SecurityInference Endpoint DiscoveryMissing AuthenticationMissing Rate LimitingAPI ExposureAI Service Fingerprinting

Plugin hooks

Custom AI Rules

Repository-local rules that extend AI Agent Evaluation without changing platform code.

AIMap

AI endpoint discovery and exposure validation for Suite 22.

Suite 23 - AI Memory & Context Security

Checks memory poisoning, persistent prompt injection, context manipulation, retrieval poisoning, long-term memory abuse, and unsafe context persistence.

Plugin ready
Memory PoisoningPersistent Prompt InjectionContext ManipulationRetrieval PoisoningLong-Term Memory AbuseUnsafe Context Persistence

Plugin hooks

OWASP Agent Memory Guard

Memory, context, and retrieval poisoning controls for Suite 23.

Suite 24 - AI Telemetry & Audit

Validates tool invocation logs, shell command tracking, file modification tracking, network call tracking, agent decision trace, and runtime audit trails.

Plugin ready
Tool Invocation LogsShell Command TrackingFile Modification TrackingNetwork Call TrackingAgent Decision TraceRuntime Audit Trail

Plugin hooks

Agent Beacon

Runtime telemetry and audit trail validation for Suite 24.

Suite 25 - AI Firewall & Runtime Protection

Validates runtime guardrails for dangerous commands, unauthorized network access, credential leakage, secret exfiltration attempts, malicious tool calls, and policy enforcement.

Plugin ready
Dangerous Command BlockingUnauthorized Network AccessCredential LeakageSecret Exfiltration AttemptsMalicious Tool CallsRuntime Policy Enforcement

Plugin hooks

Pipelock

Runtime AI firewall and policy enforcement for Suite 25.

Suite 26 - AI Privacy & Compliance

Checks AI privacy controls for PII detection, secret redaction, prompt sanitization, GDPR readiness, sensitive data leakage, and compliance validation.

Plugin ready
PII DetectionSecret RedactionPrompt SanitizationGDPR ReadinessSensitive Data LeakageCompliance Validation

Plugin hooks

Kiji Privacy Proxy

Privacy proxy, PII detection, prompt sanitization, and redaction controls for Suite 26.

Suite 27 - AI Policy & Behavior Verification

Verifies expected AI behavior against policy, drift, safety rules, goal alignment, and runtime policy expectations.

Plugin ready
Policy ComplianceExpected vs Actual BehaviorAI Drift DetectionSafety Rule ValidationGoal AlignmentRuntime Policy Verification

Plugin hooks

Praxen

AI policy, behavior, drift, and goal alignment verification for Suite 27.

Microsoft Clarity

Behavior analytics signal source for AI policy and product behavior verification.

Microsoft RAMPART

Adversarial AI behavior and safety validation signal source for Suite 27.

Agent Threat Rules

Threat rule packs for policy compliance, safety validation, and runtime behavior checks.

Plugin Registry

AI Agent Evaluation enables deeper suite behavior when these tools are found through commands, environment variables, or repository plugin manifests.

AgentGG

Optional

Optional deep AI code and agent behavior analysis for Suite 21.

Suites: ai-agent-security

Sandyaa

Optional

Authorized validation workflows for owned AI agents in Suite 21.

Suites: ai-agent-security

Custom AI Rules

Optional

Repository-local rules that extend AI Agent Evaluation without changing platform code.

Suites: ai-agent-security, ai-infrastructure-endpoint-security

AIMap

Optional

AI endpoint discovery and exposure validation for Suite 22.

Suites: ai-infrastructure-endpoint-security

OWASP Agent Memory Guard

Optional

Memory, context, and retrieval poisoning controls for Suite 23.

Suites: ai-memory-context-security

Agent Beacon

Optional

Runtime telemetry and audit trail validation for Suite 24.

Suites: ai-telemetry-audit

Pipelock

Optional

Runtime AI firewall and policy enforcement for Suite 25.

Suites: ai-firewall-runtime-protection

Kiji Privacy Proxy

Optional

Privacy proxy, PII detection, prompt sanitization, and redaction controls for Suite 26.

Suites: ai-privacy-compliance

Praxen

Optional

AI policy, behavior, drift, and goal alignment verification for Suite 27.

Suites: ai-policy-behavior-verification

Microsoft Clarity

Optional

Behavior analytics signal source for AI policy and product behavior verification.

Suites: ai-policy-behavior-verification

Microsoft RAMPART

Optional

Adversarial AI behavior and safety validation signal source for Suite 27.

Suites: ai-policy-behavior-verification

Agent Threat Rules

Optional

Threat rule packs for policy compliance, safety validation, and runtime behavior checks.

Suites: ai-policy-behavior-verification