Scan MCP server configurations for security vulnerabilities (CVEs, auth, typo-squatting, secrets)

Audit token usage and estimate costs per MCP server

Check health, latency, and reliability of MCP servers

Generate a complete security, cost, and health report for all MCP servers

Start observing AI agent tool calls to learn usage patterns for policy generation

Stop the current observation window and finalize collected data

Generate a minimal-privilege YAML policy based on observed tool call patterns

Compare observed behavior against current policy and suggest additions/removals

Get current behavior observation status and summary

Scan tool call arguments for prompt injection payloads targeting downstream AI agents

Get prompt injection detection statistics

Generate threat forecast for all configured MCP servers with 30/90/365-day projections

Detailed threat forecast for a specific server with risk factors and preemptive hardening recommendations

Get suggested preemptive policy changes based on threat forecasts

Full supply chain integrity verification with signed attestation for MCP server packages

Current trust graph state for all MCP server packages

Export Software Bill of Materials for MCP server packages

Compare current MCP server behavior against a known-good baseline to detect anomalies

Capture current server state as a known-good behavioral baseline

Revert to a previous known-good configuration snapshot

List all detected drift events

Generate auditor-ready compliance evidence bundle for a framework

Identify missing compliance controls and recommend policies

Get current compliance posture score across all frameworks

List all supported compliance frameworks

Run a full autonomous red team assessment with attack generation and policy testing

Configure periodic autonomous red team assessments

Get latest red team assessment results and recommendations

A/B test a proposed policy change against historical attack corpus

Submit an anonymized threat signature to the cross-deployment intelligence mesh

Get mesh connectivity, contribution stats, and known threat feed

Deploy an ephemeral decoy MCP server to detect adversarial probing

Get attack patterns observed by all active honeypots

Tear down a specific honeypot and retrieve captured data

List all active and destroyed honeypots with summary

Initiate an automated trust handshake with another AI agent behind Mastyf AI

View all active trust relationships and session details

Immediately terminate a trust relationship

List all registered agents in the trust registry

Get overall status of all agentic AI features including metrics, scheduler, and task queue

Compute an A+-F trust score for an MCP server across 8 security dimensions (like SSL Labs for MCP)

Scan MCP tool responses for PII, credentials, sensitive paths, and data exfiltration

Run MCP server certification (Bronze/Silver/Gold/Platinum)

List MCP servers in the local certification registry with level and expiry

Verify a server certification attestation (JWS) and level

Declare session intent and allowed tools for intent-binding enforcement

Run MCP protocol fuzzer — test defenses against malformed JSON-RPC, overflow, injection

Check SLA compliance — p50/p95 latency, error rate, circuit breaker state per tool

Execute an incident response playbook (prompt_injection, credential_leak, shell_injection)

Get agent reputation score — Trusted/Standard/Suspicious/Blocked tier with bypass rate and entropy

Analyze MCP server config and get A-F hardening grade with one-click recommendations

Detect agent-to-agent collusion patterns (recon-then-exploit, coordinated exfil, token sharing)

Explain MCP Mastyf AI policy YAML in plain English for compliance stakeholders

Convert a natural-language security goal into a draft YAML policy rule (requires approval before enforce)

Query decentralized MCP server reputation (8-dimension consensus score)

Compute cyber insurance ALE (Annualized Loss Expectancy) for an MCP server

Thompson Sampling — run Bayesian bandit trust sampling for an agent (Beta posterior, exploration/exploitation)

Contextual Bandit (LinUCB) — select optimal policy action (enforce/relax/skip) based on context

SARSA — adaptively tune rate limit, latency limit, or confidence threshold via reinforcement learning

REINFORCE — use policy gradient to select optimal fuzzer mutation strategy