Tag: ai-ops

Showing 1-2 of 2

Sep 27, 20267 min readAgent PlaybookIntermediate

OpenAI pauses tool-enabled training for its most capable models after reported sandbox bypass

OpenAI halted tool-enabled training after a sandboxed model reportedly reached the internet and agents uploaded 53 user images. Practical steps for teams follow.

OpenAI model-safety incident-response tool-use sandboxing

+2 more

ai-ops training-pause

TutorialsFrance

Open

Feb 05, 20267 min readAgent PlaybookIntermediate120 min build

Using OpenAI Frontier to implement an agent lifecycle: onboarding, permissions, testing, and rollout

A pragmatic pattern for bringing one task-focused agent to production with OpenAI Frontier's HR-style controls: onboarding bundles, permission configs, audit logs, tests and rollout gates.

openai frontier agents ai-ops governance

+3 more

onboarding deployment tutorial