Awesome Hermes Plugins › Tools & Automation

ajensenwaud/hermes-jev-plugin

⭐ 13 Python added to this list on 2026-10-12 repository created 2026-09-19

This plugin gives Hermes Agent access to TypeSafe Jev, an external "System One" decision model that evaluates typed questions against a supplied state and returns structured answers with calibrated probabilities, through four dedicated tools aimed specifically at coding-agent workflows. jev_check handles a single atomic yes/no judgment, such as deciding whether a test failure looks transient; jev_route picks one option from a fixed list, for example which subsystem owns a bug or which downstream agent should handle a task; jev_score produces an ordered rubric rating, for example severity or complexity; and jev_evaluate runs any mix of these question types in one parallel, inexpensive call when several judgments are needed on the same piece of state at once. The plugin bundles a companion skill, jev:jev-questions, that documents how to design atomic questions, prepare the state passed to the model, and gate on the returned confidence. Using it requires a TypeSafe API key, which the installer prompts for and saves to the Hermes profile's environment file, or which can be set manually; the plugin can be installed through the Hermes plugin manager or by cloning it directly into the profile's plugins directory. Configurable settings include which TypeSafe model id to use, an HTTP timeout, and a cap on how many questions a single jev_evaluate call may batch together. The README documents several concrete coding-agent recipes built on these tools: triaging a failure by asking in parallel whether it looks transient and whether it stems from a missing dependency, gating a code review by asking whether a diff would be blocked by a senior engineer or introduces a security risk, and routing a task description to the right option among a fixed set of choices.

https://github.com/ajensenwaud/hermes-jev-plugin

decision-modeltoolsstructured-judgmentexternal-apihermes-agent