AI that acts on ground truth.
Most AI grades identity against role definitions — intent, not reality. SimpliKeys feeds detection the live truth of every session, identity, and action, human and non-human.
AI-driven threat detection.
sk-control — threat-detection
live$ simplikeys detect --post-exploitation --c2 --contextual
▸ credential-attacksKerberoasting · Password Spraying · Pass-the-Hash (Mimikatz) · LDAP Relay · DCSync · Golden Ticket · LLMNR Poisoning
▸ operator-behaviorPrivilege Escalation · Maintaining Access · Data Gathering · Pivoting · Backdoor Creation · Command & Control · Evasion
▸ c2-frameworksMythic · Sliver C2 · Octopus C2 · BadPotato · Deimos C2 · Nimplant — others added as patterns emerge
▸ contextual-riskimpossible-travel detection · access timing · multiple concurrent sessions · behavioral deviation from baseline
$ ▋
02
The kill switch your AI doesn't have.
Agents do the work they're scoped for — and nothing more. Run one and watch privilege escalation and self-modification get stopped at runtime.
agent: support-bot · runtime guard ONmockup
Click an action above to make the agent attempt it.
Inside the AI engine.
AI Threat Detection
Behavioral analytics across every identity and session — credential attacks, C2 frameworks, post-exploitation, and contextual risk.
Non-Human & Agent Identity
API keys, tokens, certificates, and AI agents inventoried and governed — with runtime control over what agents can do.
Behavioral Ground Truth
Detection grades against what identities actually do, not how roles are defined — so over-privileged accounts can’t hide.