It's well known that agents hack benchmark rewards. The usual response is a patch for each task that gets exploited. In a study of 456 adjudicated trajectories from more than 31,000 public agent runs, 69% contained at least one reward-hacking episode. Most of the exploits appeared mid-run after legitimate work. BenchShield models each evaluation as a finite set of reward-relevant events. A static taint analysis finds hack paths from the task package before any run. A runtime pass then uses evidence from the benchmark infrastructure to decide whether the agent actually used one. On Terminal-Bench 3, SkillsBench and ClawsBench, the static pass recovers 77 to 100% of exploit chains, against 23 to 94% for an agentic scanner, at up to 65% lower cost. Runtime detection reaches 96% accuracy, against 36% for an LLM reading the transcript. Paper:

Finding an AI agent is easy. Deciding whether it fits a real task is harder. This guide shows how turns public agent data and live audit evidence into a practical selection workflow.

An Agent Card shows what an AI agent claims. checks what actually works. See how to audit a public agent and review its protocols, capabilities, security checks, and evidence.