Skip to content

Jev directory

Pi Skimmer

jon-devlapaz's MIT-licensed Pi coding-agent extension that filters text and JSON through Jev with explicit per-session consent, bounded requests, owner-only recovery files, and unjudged material preserved.
Stars
2
Language
TypeScript
License
MIT
Reading time
3 min
Published by
Community
Author
jon-devlapaz
Last verified
Oct 2026

Highlights

  • Two Jev-backed tools for the Pi coding agent: skim_text keeps relevant chunks with line ranges, filter_items keeps relevant JSON records.
  • Consent is separate from /typesafe enable, scoped to the session and working directory, and revoked on reload; headless use needs PI_SKIMMER_ENABLED=1.
  • Budget guards: 100 attempted requests per extension session, a shared estimated $1 daily spending cap, a 10 s request timeout and a 30 s tool timeout.
  • skim_text defaults to 2,000-character chunks at 10% overlap and a 0.5 relevance threshold; filter_items takes 1-256 items on literal top-level fields.
  • Input over 256,000 UTF-8 JSON bytes fails rather than truncating, and unjudged items are retained with warnings instead of being dropped.

Watch out

MIT-licensed. Needs Pi 1.2.0+ and Node 22.19+; the npm package is promised after release, with git install available now. Field selection is not output redaction, spending estimates are not billing guarantees, and relevance scores are probabilistic, so the README says to read originals before critical decisions. It is a native Pi extension, not a wrapper around the Skimmer CLI.

Questions about this entry

Short answers drawn from the linked project or page.

From the Jev guides

Original write-ups on what this entry is for.

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai