Unit 42 AI Security · September 18, 2026

AgentCore Harness: shell access can expose credentials after vault retrieval

Why it matters

Unit 42 demonstrates prompt injection reaching plaintext integration credentials through AgentCore Harness shell access in a permissive test configuration. AWS classified the report as informative under shared responsibility, pointing to tool scoping and egress controls.

My takeaway: Disable unnecessary shell tools, separate credential-bearing processes from agent-executed code, scope downstream identities, and test outbound controls. Encryption in a vault does not isolate credentials once a runtime can read them.
Keep exploring

More curated notes connected through Agent Security and Prompt Injection.

OpenAI News · framework

Pacing model development in an era of cyber-critical capabilities

OpenAI says preliminary evidence that Astra may meet its Critical cybersecurity threshold led it to pause frontier reinforcement-learning work for two weeks and keep its largest planned run on hold. New safeguards include stronger workload and network isolation, continuous boundary testing, token-level monitoring that escalates suspicious tool activity, and broader alignment checks for deception, reward hacking, and unauthorized access.

OpenAI News · framework

A blueprint for democratic governance of frontier AI

OpenAI proposes a three-part U.S. frontier-AI governance model: harmonize emerging state safety laws into a federal baseline, strengthen CAISI as an evaluation and standards institution, and coordinate a broader resilience program. Proposed controls include severe-risk evaluations, transparency reports, independent audits, safety-incident reporting, model-weight security, whistleblower protection, and periodic technical assessments.

OpenAI News · framework

OpenAI’s Frontier Governance Framework

OpenAI's 22-page Frontier Governance Framework maps its frontier-model processes to California's Transparency in Frontier AI Act and the EU AI Act's general-purpose AI code. It documents lifecycle risk assessment, cyber-offense and other risk tiers, mitigation and residual-risk decisions, critical-incident handling, security risk management, model reporting, external review, responsibility allocation, and change control.