Agents

Benchmark

Laminar’s RL-trained model finds agent-trace errors at 23 times lower cost

Laminar says its flow-1 model matches GPT-6-sol on its trace-error benchmark while costing 23 times less, making full production-trace analysis more affordable. Teams evaluating agent observability should examine how the benchmark and cost comparison hold up.

Original · news.ycombinator.com 13 min read HN 58 pts · 27 comments

Guide

Harness Engineering 101 builds an agent from a 300-line Python loop

The series grows a roughly 300-line Python agent across 16 chapters, framing the harness as the program that builds, maintains, and protects the conversation sent to a stateless LLM. That grounding gives engineers a practical way to reason about agent behavior and failure modes.

Original · github.com 3 min read HN 132 pts · 64 comments

Vulnerability

Muse conversation links exposed users’ personal and business secrets

Researchers found Muse’s share URLs were easy to discover, potentially exposing conversations and other shared content; Meta suspended the feature for review. Agent builders should treat link generation and access controls as security-critical design choices.

Original · arstechnica.com 4 min read HN 67 pts · 13 comments

AI agents carried out attacks on more than 100 websites with little human input

Norforge says AI agents autonomously carried out attacks on more than 100 websites, with Reuters identifying John Deere, SharkNinja and Cummins among the victims. The incident makes agent autonomy and security controls an immediate operational concern.

Original · theverge.com 3 min read HN 59 pts · 3 comments

Paper

MMEnterprise-Bench tests agents on conflicting, evolving workplace records

The benchmark contains 1,613 questions testing evidence reasoning, conflict resolution, and unanswerability; evaluated models struggle with cross-source and temporal conflicts. Larger retrieval budgets improve recall but not consistently answer accuracy, a useful warning for enterprise-agent design.

Original · huggingface.co by Seohyun Lee et al.15 upvotes Code · 1 stars

Paper

S-Agent 2.0 improves spatial reasoning with a coding-agent perception toolkit

A command-line toolkit gives a coding agent persistent scene memory and plain-language interpretations of geometric readings; it improves Qwen3.8-27B by 13.9 and 20.5 points on two benchmarks. The results show how tool interfaces can turn perception outputs into usable agent evidence.

Original · huggingface.co by Rui Gong et al.7 upvotes Code

Benchmark

ALTK-Evolve reaches 67.5% Pass⁴ on τ²-bench retail with learned memory

ALTK-Evolve uses learned memory with GPT-4.1 to reach 67.5% Pass⁴ on τ²-bench retail, above every listed retail entry. The result suggests memory can substantially improve a basic customer-service agent without changing model weights.

Original · huggingface.co 10 min read

Essay

Agent monitors watch service telemetry for anomalies fixed alerts may miss

The proposed “superfans” each monitor one system component using telemetry and logs, looking for subtle or novel issues beyond traditional alerts. This operational pattern could extend monitoring coverage, though the excerpt gives no measured results.

Original · benjamincongdon.me 9 min read HN 118 pts · 23 comments

Guide

MCP Apps add interactive, synchronized workspaces to Gemini Enterprise chats

MCP Apps pair a tool with a live UI in a sandboxed iframe, letting users make local edits without another model turn while keeping the assistant synchronized. The guide outlines a practical pattern for richer tool interactions in enterprise agents.

Original · cloud.google.com by Dave Wang5 min read