SignalTech/AI
Doc-foundry – AI writes Security detection rules, then proves they work
AI generates security detection rules and validates their effectiveness, signaling practical AI security tooling.
Context
What happened: Doc-foundry turns a threat such as a CVE, CISA KEV entry, incident writeup, or ATT&CK technique into a validated first-version Sigma detection rule, shipped with a threat report and a proof bundle.
Key numbers: The proof bundle includes a coverage label (real-coverage/brittle/broken) with FP/FN numbers and a reproducibility hash.
Why it matters: The content states the proof bundle is what other tools do not produce: a model can write a Sigma rule but cannot hand over proof that it works, an adversary that evades the weak version, and a record of how the rule was hardened.
What changed
Doc-foundry turns a threat such as a CVE, CISA KEV entry, incident writeup, or ATT&CK technique into a validated first-version Sigma detection rule, shipped with a threat report and a proof bundle.
Why it matters
The content states the proof bundle is what other tools do not produce: a model can write a Sigma rule but cannot hand over proof that it works, an adversary that evades the weak version, and a record of how the rule was hardened.