Security correlation
How Hanu joins typed signals from cloud, code, identity and AI workloads into a single compound risk with a deterministic score and a full evidence trail.
security-correlation
Seven attack classes, one signal bus
These are not product modules sold separately. They are the same collectors and the same correlation engine, described from the angle of the attack each one answers — which is how the work actually arrives.
Supply chain security
Poisoned releases, tampered CI workflows and credentials exposed in git history — detected separately, joined into one compound risk.
supply-chain-security
CI/CD security
Workflow tampering, unpinned actions, provenance gaps and release freshness as first-class signals, with the build-log evidence attached to each finding.
cicd-security
LLM & AI workload security
Detect hijacked credentials being spent against model APIs, and find the AI workloads missing from your inventory — with no LLM in the detection path.
llm-security
Agent & MCP security
Agent identities, tool permissions and MCP servers are infrastructure. Track what an agent can reach, what it actually reached, and which of its tools nobody owns.
ai-agent-security
Non-human identity security
Service accounts, CI credentials, keys and OIDC trust relationships, inventoried from the cloud APIs and scored on use, age and reach.
non-human-identity
Attack path analysis
An exposed workload, a role that can assume another, a datastore behind both. Alone they are low findings; joined in one graph, the chain is critical.
attack-path-analysis
External attack surface
Certificate transparency as the inventory: subdomains you forgot, CNAMEs pointing at nothing, and certificates chaining through a compromised CA. With the limits stated.
external-attack-surface
Compliance mapping
Thirty controls across SOC 2, ISO 27001, NIST 800-53, PCI-DSS, HIPAA and GDPR, each scored from live scan findings. What that is good for, and what it is not.
compliance-mapping