AI Engineer · November 5, 2026

Delegated agent access: narrow tokens at each hop and reacquire on execution

Delegated agent access: narrow tokens at each hop and reacquire on execution video thumbnail
Why it matters

A Keycard and Temporal workshop demonstrates an expense agent moving from shared API credentials to explicit agent and user identities. Token exchange narrows access across the agent, MCP server and ledger API, while audit records preserve who acted for whom. Durable execution resumes recorded work, and an execution interceptor obtains credentials without storing them in workflow history. This is an instructional demo, not a general outage or authorization guarantee.

My takeaway: Test resource audience, scopes, delegation identity and self-approval denial at every hop. Keep bearer tokens out of durable history and acquire them when protected activities run. Exercise restart, expired-token and revoked-access cases, and use idempotent downstream operations to handle uncertain retry outcomes.
Keep exploring

More curated notes connected through Agent Security and AI Engineering.

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.

OWASP GenAI Security Project · guide

OWASP Top 10 for Agentic Applications for 2026

OWASP's community guide organizes agentic-system risk into ten categories, including goal hijacking, tool misuse, identity and privilege abuse, memory poisoning, insecure inter-agent communication, cascading failures, and rogue-agent behavior. It provides a shared taxonomy and mitigation starting point rather than a certification checklist or evidence that a deployed system is secure.

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.