StateM: Reaching 95.3% Raw Accuracy, or a \$15 Frontier Run, on Terminal-Bench 2.1 via Harness Scaling · HackerPapers discussions

StateM: Reaching 95.3% Raw Accuracy, or a \$15 Frontier Run, on Terminal-Bench 2.1 via Harness Scaling — mentioned by johntrob14

Mon, 24 Aug 2026 18:21:49 GMT

I started using this tool called StateM for my agents lately. It is an open-source control layer for long-horizon agents, and I felt it was worth sharing here. The basic idea is to move parts of agent control from the prompt or agent context to an externalized layer, making them explicit and enforceable. The interesting thing is what they call an “agent owning the control”: StateM uses YAML-defined runbooks and a CLI runtime to provide things like stateful checks, phase-specific context, blocking transitions, and persistent workflow state. It’s very easy to implement - my agent is already…

Paper and discussion history. A mention is not an endorsement.

StateM: Reaching 95.3% Raw Accuracy, or a \$15 Frontier Run, on Terminal-Bench 2.1 via Harness Scaling — mentioned by johntrob14

Sat, 22 Aug 2026 14:07:09 GMT

Hi all, I came across an interesting project called StateM. It is an open-source control layer for long-horizon agents, and I felt it was worth sharing here. The basic idea is to move parts of agent control from the prompt or agent context to an externalized layer, making them explicit and enforceable. The interesting thing is what they call an “agent owning the control”: StateM uses YAML-defined runbooks and a CLI runtime to provide things like stateful checks, phase-specific context, blocking transitions, and persistent workflow state. The developing speed is great -- my agent is already…

Paper and discussion history. A mention is not an endorsement.

StateM: Reaching 95.3% Raw Accuracy, or a \$15 Frontier Run, on Terminal-Bench 2.1 via Harness Scaling — HN submission

Wed, 19 Aug 2026 06:15:15 GMT

StateM: Reaching 95.3% on Terminal Bench 2.1

3 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.