About
AI Security Index is a machine-swept, human-verified feed of AI security news: CVEs, incidents, jailbreak and prompt-injection research, red-team reports and supply-chain findings.
The policy
- ≤72h old, or older if first-reported with a primary source.
- Primary source required: advisory, vendor post, paper, CVE, research repo.
- A source that returns 403 and cannot be corroborated is dropped — and the drop is logged with the reason.
- Nothing is hallucinated. If a URL or claim cannot be verified, the item is dropped.
- Summaries are drafted by a local model, reviewed by a human; CRITICAL always is.
Runs as an independent project. DevMinimum is the one-person studio behind it — that is the only promotion this site carries.