AI Engineer · October 6, 2026

Improve agent skills through traced failures and protected evaluation criteria

video Model Evaluation AI Engineering Agent Security

Improve agent skills through traced failures and protected evaluation criteria video thumbnail

Why it matters

Marc Klingen describes an improvement loop that turns observed failures into evaluation cases, scores candidate changes, and sends proposed fixes for human review. The useful boundary is that an optimizing agent should not silently redefine its own success criteria. A companion Langfuse walkthrough provides a concrete skill-evaluation setup: start each run from the same repository state, capture tool calls and edits, and inspect failures such as missing configuration or invented CLI arguments. These examples demonstrate how to investigate a regression; their local outcomes are not proof that an agent will improve safely or generalize to new tasks.

My takeaway: Freeze a held-out dataset and scoring rules before editing a skill. Compare error counts, task correctness, retries, and total tool calls on clean fixtures; require separate review for evaluator changes and for the proposed production change.

Keep exploring

More curated notes connected through Model Evaluation 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.