Jevgrep: a semantic code search CLI that helps coding agents find code by behavior
blaizedsouza · x · 2026-10-05
A new CLI tool, jevgrep, gives coding agents semantic code search for the case where you know what a behavior does but not where it lives. Instead of matching exact strings, filenames, or symbols, you can ask repo-level questions like "where is authentication checked before a request reaches a handler?"
- Jevgrep explores the repository and uses an LLM (Jev) to judge relevance across folders, files, and declarations
- It returns relevant file locations with verbatim source excerpts and line references
- The coding agent then does the actual work: reading the source, making changes, and running tests — the LLM is only used in the retrieval layer
The tweet also references an explainer article on how agent runs are mostly small decisions rather than pure reasoning.
More from coding & agent
- New Paper IGP-Bench Teaches AI Agents When to Stop Chasing a Wrong Idea — _sathvikr · 2026-10-05
- AI decompilation era: PS5 is 80% ported to PC and 'closed source' is ending — almmaasoglu · 2026-10-05
- Claude Code ignores steering messages, Codex always acknowledges them — Angaisb_ · 2026-10-05
- Dad of 3 built a chat plugin to control Ableton in plain language, no coding — _AustinCalvert_ · 2026-10-05
- Engineer Builds Social Network for AI Optimists in 2 Weeks via Voice Prompting — RoyalCultural · 2026-10-05
- mitsuhiko explains why Pi carries extra code to support multiple Python installers — cnakazawa · 2026-10-05