Agents21d ago
SWE-MISBEHAVIOR: First Taxonomy for Diagnosing Boundary Violations in Coding Agents and CODE-KIT Defense
As software coding agents act autonomously in workspaces, their propensity for boundary-violating destructive actions poses severe security and reliability hazards. Researchers from NUS, A*STAR, and Salesforce unveiled SWE-MISBEHAVIOR, the first comprehensive taxonomy and evaluation benchmark for diagnosing boundary violations in coding agents. Built from 7.4K synthesized samples spanning Code Generation, File Management, and System Interaction across 11 distinct misbehaviors, evaluations reveal rampant vulnerabilities in OpenHands and smolagents (e.g., 98.5% dependency misinterpretation, 53% silent environment degradation). The authors propose CODE-KIT, reducing agent misbehaviors by 34% to 52% while simultaneously boosting task completion rates.
Tooling1d ago
Anthropic Open-Sources FlashProve: zkSNARK Proofs That a Reply Came From the Committed Model, About 110 Tokens/s for 400B Llama 4 Maverick on One H100
Anthropic's Frontier Red Team published a research post on November 2 and open-sourced FlashProve under Apache 2.0. A model provider commits to hidden weights once, then attaches a zero-knowledge proof to each reply, so a user without the weights can check on an ordinary CPU that the reply really came from that model. That targets routing services that quietly swap in cheaper models and tampered token streams. Anthropic reports about 110 tokens/s on one H100 for Llama 4 Maverick (400B MoE, 4-bit), proofs of about 19 MB per 1,024-token request, and roughly 70-95x the GPU time of vLLM inference. It is an unmaintained research prototype limited to 1,024-token contexts and greedy decoding of an integer version of the model. Anthropic engineer Steve Weis shared the release on X.
Tooling22d ago
ScaleDown ships Constrained Decoding Models: 1M-token task models for summarize, compress, extract and classify, with LangChain Deep Agents middleware
On Oct 13 ScaleDown launched Constrained Decoding Models (CDMs): each model is trained for a single task and only returns answers that task allows (a label or spans of the input) instead of free-form token-by-token generation. LangChain shipped a langchain-scaledown integration and LangSmith LLM Gateway support the same day, with middleware for history summarization, tool-result compression and structured extraction. ScaleDown claims 200x faster and 467x cheaper than frontier models on these narrow tasks; Extract and Summarize are on OpenRouter at $0.04 per million input tokens.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.