#259259th of 455 in Skills, plugins and prompts
hermes-jev-skills
Skills and Hermes plugin that hand small agent decisions to TypeSafe's Jev
- Stars
- 1.3k
- License
- MIT
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- Python
3 on 15 Sept 20261,337 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Thirteen agent-agnostic SKILL.md files plus a Hermes plugin that send small decisions to Jev, TypeSafe's cloud decision model: per-turn model routing, skill selection, search ranking, memory filtering, web-result injection screening, mailbox sorting, and GUI or browser action choice. Jev never writes text; it returns calibrated confidences, and every feature fails open to current behavior. A stdlib-only Python installer sets it up for Hermes, Claude Code and Codex.
Who it is for: Hermes, Claude Code and Codex users cutting routine frontier-model spend
Strengths
- Shadow mode logs routing decisions without switching models, so you can verify first
- Fails open on no key, timeout or low confidence; routing budget is 2.5 s
- Installer is stdlib-only Python 3.9+, with --check dry run and --uninstall
- API key goes to the OS secret store via a local one-time page, not the agent
Weaknesses
- Requires a TypeSafe API key and cloud calls; no self-hosted Jev is mentioned
- Default routing sends the redacted turn text off-machine unless the profile is private
- Full integration is Hermes-only; other agents get just the skill files
- Web screening recall is in-distribution evidence and the README says not portable
What it needs
- no GPU
- Needs TypeSafe Jev API key, Python 3.9+, models.dev catalog
- Models: Any model reachable through Hermes providers (e.g. OpenRouter), Claude Code, Codex
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Skills, plugins and prompts
See all 455| Rank | Project | Score |
|---|---|---|
| 1 | claude-memPersistent cross-session memory plugin for Claude Code and other coding agents | 97 out of 100 |
| 2 | impeccableDesign skill with 24 slash commands for AI coding agents' frontend output | 97 out of 100 |
| 3 | gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles | 96 out of 100 |
| #4 | rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex | 96 out of 100 |
| #5 | cavemanShortens coding-agent replies and compresses tool output before the model reads it | 95 out of 100 |