#9292nd of 98 in Skills, plugins and rules

fast-jev-compaction

Claude Code plugin that prunes stale tool calls instead of summarizing history

Stars: 7.6k
License: MIT
Last commit: Sep 2026
Language: TypeScript

Missed the latest GitHub check; these facts may be older.

Overview

Replaces Claude Code's compaction summary with a pruned copy of the original transcript. Each tool call and result is scored by the Jev model through TypeSafe's API, and stale ones are dropped or truncated while user and assistant text stays verbatim. It also works as an npm library, and it falls back to the built-in summary on errors or small reductions.

Who it is for: Claude Code users who want lossless context compaction and accept a hosted API

Strengths

Never rewrites content: kept messages stay verbatim, only tool calls and results are removed
Falls back to Claude Code's built-in summary on errors or insufficient reduction
Also exported as an npm library with a pluggable JevAsker transport and building blocks
Tunable keepThreshold, preserveRecentMessages, token ceilings and truncateHeadChars

Weaknesses

Requires a TypeSafe API key and sends the conversation to an external Jev endpoint
Needs Claude Code 2.1.274+ with the early-access function hooks flag enabled
Text messages are never shortened, so long text-heavy sessions may reduce little
Token counts are estimates, and histories that cannot fit the state ceiling throw

What it needs

no GPU
Needs TypeSafe API key (Jev model, api.typesafe.ai), Claude Code 2.1.274+ with CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1, Node.js and npm
Models: jev-latest

Also in Skills, plugins and rules

See all 98

Also in Skills, plugins and rules
Rank	Project	Score
1	skillsAgent skills for planning, TDD, debugging and code review291.3k starsMIT	94 out of 100
2	eccAgents, skills, hooks and rules bundle for Claude Code, Codex and other harnesses279.5k starsMIT	94 out of 100
3	cavemanCuts token use in coding agents with terse replies and compressed inputs Live demo ↗ (opens in a new tab)111.9k starsApache-2.0	94 out of 100
#4	impeccableDesign skill with 24 commands and a detector for AI-generated frontends Live demo ↗ (opens in a new tab)81.2k starsApache-2.0	93 out of 100
#5	ponytailSkill prompt that makes coding agents write less code and add tests163.6k starsMIT	92 out of 100

How to choose

Check which agent it targets and whether it works with the version you run.
Read what it adds to every prompt; big rule sets cost tokens on each call.
Prefer add-ons you can turn off per project.

Health

Releases: Not checked
Top contributor: Not checked
Security advisories: Not checked
Package advisories: Not checked
OpenSSF Scorecard: Not checked

Not read for this project yet: each refresh reads the facts of a few projects. Bots and coding-agent accounts are left out of the commit counts. Not part of the score: how we check.

License in plain words

MIT Permissive

Use it at work: Yes
Change it and rebrand it: Yes
Sell it or host it for others: Yes

Your part: Keep the copyright and license notice in copies you ship.

A plain summary, not legal advice. Read the license on GitHub ↗ (opens in a new tab)

#92 of 98 in Skills, plugins and rules on Archestack, as of 15 Oct 2026.

In your README

The badge updates twice a day.

[![Archestack](https://archestack.com/badge/tamaratran/fast-jev-compaction.svg)](https://archestack.com/best-of-vibe-coding/projects/fast-jev-compaction/)

How badges work

On your site or blog

A card with the rank, score and signals, in light or dark to match the reader.

<iframe src="https://archestack.com/embed/best-of-vibe-coding/fast-jev-compaction/" title="fast-jev-compaction on Archestack" width="400" height="270" style="border:0;max-width:100%" loading="lazy"></iframe>

Preview the card

Best of Vibe Coding on GitHub

Each review comes from the project's README and facts we check on GitHub. The fix form opens with fast-jev-compaction filled in.