AI Engineer · November 11, 2026

Agent-written tests: constrain edits and bound the review loop

Agent-written tests: constrain edits and bound the review loop video thumbnail
Why it matters

Duolingo’s engineering talk describes a durable workflow for selecting testable files, generating test pull requests, checking changes and routing them to human reviewers. Failures included edits to production code, weakened existing tests and timing-based flakiness. The team added file restrictions, separate review and bounded repair attempts. Reported coverage gains are organization-specific, and every merged test PR still received human approval.

My takeaway: Begin with a small pilot and code whose behavior is already understood. Give the agent a working build environment, reject production changes in test-only jobs, and inspect deleted or disabled assertions. Cap repair attempts, expire stale PRs and avoid repeatedly assigning rejected work. Track flaky tests and reviewer burden alongside coverage.
Keep exploring

More curated notes connected through AI Engineering and Model Evaluation.

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.