Together AI reports prompt leak after Dragos canary experiment
Together AI disclosed that a coding agent exposed a credential for stored prompt samples, according to The Defiant. The report states that five credentials later tested by the intruder were canaries from a Dragos experiment.
Why “Unverified”
- Reported by The Defiant alone, with no second source or primary confirmation yet
1 independent source group across 1 post · evidence score 52/100. How labels work
What the sources claim
-
Together AI disclosed that a coding agent exposed a credential for stored prompt samples.
“Together says a coding agent exposed a credential for stored prompt samples.”
The Defiant -
Five credentials later tested by the intruder were canaries from a Dragos experiment.
“Five credentials later tested by the intruder were canaries from a Dragos experiment.”
The Defiant
Sources
More in Hacks & Security
Embed this page
Summary and claims extracted by AI with prompt analyze-story@5 on 2026-11-21 23:40 UTC. Labels are computed from the sources, not by the model.