Small decisions. Interesting possibilities.Submit contentSubmit

jev-skill-suggester

用 TypeSafe Jev 推荐已安装 Skill / Bounded installed-skill recommendations with TypeSafe Jev. Python CLI, Codex skill, bilingual docs and live examples.

Added to Jevfast

How it uses Jev

In Jev mode, the CLI first narrows skills using descriptions, then asks Jev to rank candidate descriptions with Choice and independently judge group usefulness with Noul. It reviews candidate text snippets and uses a final Choice plus per-candidate fit Noul checks; local mode does not claim a Jev recommendation.

What Jev decides

Illustrative Jev requestExample answers · not a recorded Jev response · Source ↗
Question 1 · needed
YOUR APP
INSTRUCTION

Does any skill in this roster provide a specialized procedure useful for this request?

STATE

User task and roster of skill IDs, names, descriptions and hashes.

JEV · NOUL
YesNo

App workflow

  1. Read local skill metadata

    Collect installed skill names and descriptions, while honoring a skill the user explicitly named through local lookup.

  2. Filter and evaluate candidates

    In Jev mode, rank descriptions, assess whether candidate groups are useful, and inspect limited candidate text snippets.

  3. Return a bounded recommendation

    Apply the documented exploratory fit and confidence thresholds, or return none, uncertain, or local-only status.

Why it is interesting

It explicitly supports no-skill and uncertain outcomes, prioritizes a user's named skill through local lookup, and treats Jev's typed decision as recommendation input rather than execution authorization.