Skip to content
Research

Measured, not marketed.

Reproducible detection-accuracy benchmarks, the Shadow AI Risk Index scoring methodology, and our flagship State of MCP Security landscape report.

We publish measured results and original analysis — not marketing claims. Everything here is reproducible: we tell you the method, then the numbers.

Detection-accuracy benchmark

Security tools are only as good as what they actually catch. We benchmark our own detection the honest way: plant a fixture of known issues alongside clean controls, run the real collector/adapter against it, and measure recall (planted issues detected) and false positives (findings in clean files). No estimates.

CapabilityWhat's measuredRecallFalse positives
Endpoint sensor — discoveryAPI keys, MCP configs, AI CLIs/SDKs, local skills across a planted $HOME100%0
Source Code Analysis (SAST)Planted vulns vs clean controls (semgrep pack)100%0
Secret & Credential ScanAWS/GitHub/Slack/Stripe/OpenAI keys + private keys100%0
Dependency / Supply-chain (SBOM)Vulnerable + typosquatted packages vs safe pins100%0
Container & IaC postureMisconfigurations vs hardened baselines100%0
AI-BOM & License/ProvenanceModel files, pickles, copyleft licences, unpinned revisions100%0

Each capability was tuned iteratively — baseline → measure → fix → re-measure — and the benchmark caught real bugs along the way (e.g. a missed ~/.mcp.json at the home root, an unscanned .pem extension). Full methodology and per-iteration results are in the reports:

Sensor benchmark (PDF) ↓

The Shadow AI Risk Index — methodology

A repeatable way to score an organisation's agentic exposure. The Index combines five weighted surfaces into a 0–100 score, so posture is comparable over time and across business units:

SurfaceWhat it captures
Unsanctioned AIAgents, CLIs, SDKs and apps running without approval, weighted by reach.
MCP & tool exposureMCP servers and the blast-radius of the tools they expose (file, network, credential access).
Secret exposureAPI keys and tokens reachable by AI processes, weighted by provider and scope.
Egress & data flowSensitive data paths to external AI endpoints; DLP coverage.
Governance gapAttribution, offboarding, and framework-control failures (OWASP LLM / NIST AI RMF / EU AI Act).

The Index is the scoring model behind the platform's posture score. We're publishing the framework openly so teams can reason about agentic risk consistently — the full specification ships with the flagship report below.

The State of MCP Security 2026 In progress

Our flagship landscape report — the MCP threat taxonomy (tool poisoning, confused deputy, slopsquatting, over-scoped tools, tool-output injection), what the public advisory record shows, the Shadow AI Risk Index specification, and our measured detection results. It draws on public advisories and incident reports and our own benchmarks — no invented telemetry. Get notified on release →

Reproduce it yourself. Our benchmark fixtures and method are documented — request them with an assessment enquiry and we'll walk your team through the harness.