agent-diagnostics is an open-source framework for analyzing and annotating the behavior and reliability of coding agents. It provides tools for benchmarking, taxonomy creation, and structured evaluation, helping researchers and developers understand agent performance.
agent-diagnostics sits in PulseGate's LLM eval & observability category. It focuses on understanding and diagnosing why coding agents succeed or fail through structured analysis. agent-diagnostics is an open-source project aimed at AI researchers and developers working with coding agents. The project is open source (Apache-2.0). It ships for the command line.
Behind agent-diagnostics is sjarmak, and it first shipped in 2026. The project is developed in the open on GitHub with 146 commits in the last 90 days. Among its 5 catalogued features are behavioral annotation, reliability analysis, and taxonomy framework.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do