Jev for Claude Code, Codex, and Coding Agents

tamaratran/fast-jev-compaction

Claude Code plugin that replaces the compaction summary with Jev decisions. Every tool call and result is scored; stale ones are dropped or truncated, and everything kept stays verbatim.

thruwire/foreman

Puts Jev as a fast supervisor above slower coding agents such as Codex, starting from a ticket, spec, or bug report.

devagrawal09/jev-review

Staged code-review workflow for JavaScript and TypeScript with a local dashboard. Jev screens correctness, security, reliability, compatibility, and test risk, then scores severity and suggests a reviewer, with no generative model involved.

qybaihe/mu

Coding agent built on pi with a judgement kernel deciding the routine calls. Five languages.

TianyuCodings/JevHarness

Claude Code and Codex skill in which an LLM writes a task-specific harness that turns observations into typed Jev questions, then revises it from rewards and full traces. No licence yet.

lakeday-org/perch

Semantic linter that reads each method together with its callers and callees before asking about it. Rules are sentences in a YAML file.

codejunkie99/keel

Local-first macOS workspace around the coding agents you already run: a local Laya selector, or an optional hosted one, proposes the route and the host checks it before Keel applies it.

0xNatoshi/jev-codex-router

Picks the model, reasoning depth, and speed mode for every Codex turn based on how hard Jev judges it to be. The author reports about 60% lower cost when replaying 237 real turns.

NiazMorshed2007/jev-review

Local-first MCP server for continuous software-quality review by coding agents.

kunchenguid/compact-adviser

Agent plugin that asks Jev whether the session is at a safe point to /compact, and can run it automatically on Pi and Claude Code.

mitkox/esf

Self-hosted software factory: coding agents in microVMs, changes verified, and the patch and execution evidence kept, with Temporal coordinating the workflow.

bgts-ai-org/bgts-context-engine

Deterministic code-graph context for coding agents on PostgreSQL, Apache AGE and pgvector, over MCP and REST; Jev or a local decider model can optionally re-select the ranked symbols, failing open to the plain ranking.

tamaratran/jev-pruner

Claude Code plugin that uses Jev to trim long Bash output before it reaches the model, leaving errors, source code, and structured output untouched.

y0usaf/pi-jev

Jev as a decision layer for the Pi coding agent in three places: a gate that judges bash, write and edit calls before they run, an output judge that reads what a bash call printed, and a tool the model can call directly.

supercorp-ai/supercov

Coverage, security and code quality for coding agents. Jev answers twelve quality and twelve security questions about each source file, and the CLI ranks the files so the agent knows what to fix first.

egma-ai/jev-code-reviewer

Reviews your own agent's pull requests on your machine and posts nothing to GitHub, prioritising where a human should look rather than restating the diff.

Dicklesworthstone/skillranker

Rust CLI that ranks which agent skills fit the next step from live session context, with Claude Code hooks.

Avinash-jetwani/jevmem

Automatic project memory for Claude Code, Cursor and Codex, with typed decisions choosing what is worth remembering.

joetawil7/first-pass

Claude Code plugin with rules and checks that make the agent look around a change before calling it done; an optional Jev judge decides whether a review finding is real harm and what proof a small fix needs.

codegirl-007/jevlint

Extracts code units with Tree-sitter and asks whether each passes a rule written in plain language; prebuilt binaries. Unrelated to the JevLint above.

mizchi/jev-lint

Checks whether a function does what its name says, whether a comment is still true, and whether a test verifies what it claims, with a cutoff per rule.

devagrawal09/stanley-code

Bounded Jev workflows for coding agents (formerly jev-code).

GhalebDweikat/winnow

Context sieve for Claude Code. Jev judges each tool result (Read, Bash, Grep output) for relevance before it enters the context window.

Sev7eEn7/dsh-sieve

DeepSeek Harness plugin that filters tool output, prunes old context and discloses skills progressively, with Jev or Laya as the judge of what can go; it never uses the session's own model for that. English and Chinese.

wobsoriano/oxlint-plugin-jev

Oxlint rules written as plain-English questions about a function, call, JSX element or file, reported when the yes-probability clears your cutoff.

jonathanavis96/jev-kit

Claude Code bundle: a tool-call guard, sub-agent sizing, and Jev wired into search, browsing and review.

ericjuta/omp-optchat

OMP plugin for one endless chat kept in a binary tree of summaries, with five Jev gates through OMP's judge role: standing rules, history prefetch, compression hints, tool-call screening and recall. The gates send judged text to the judge's provider. No licence file.

EliaAlberti/jev-rules

Scores your standing Claude Code instructions against each prompt and delivers only the ones Jev picks, once per session rather than per message. Ships a pane that shows which rules were chosen.

Rikinshah787/dotpals

A desktop pal that watches Claude Code, Codex or any agent and says in plain words which files changed, which commands ran and what failed. Runs locally.

nahid-sparktales/agent-dispatcher

Routes a Claude Code or Codex task to one of 27 specialist roles and defines what evidence will count as done.

keeltrace/hermes-nerve

Supervisory layer for Hermes agents that adds typed decisions, ranking and verification asynchronously, with Jev authoritative and an open model shadowing it.

Oqura-ai/optim-jev

Claude Code plugin that uses Jev to drop stale tool output during compaction, route file edits to cheaper models and pick a short list of relevant skills per prompt.

danielmiessler/Deeds

CLI and Claude Code plugin that reads each commit's diff and counts capabilities gained, fixes and upkeep, ignoring commit messages. Needs only a Jev key; only diffs and file paths are sent.

IAmUnbounded/shed

Message the coding agents on your Mac from your phone, with a local Laya model choosing which installed harness and model handles each new task. No TypeSafe account or key.

korallis/agent-stack

Turns one Linux machine into a team of coding agents that plans, writes tests first, builds, uses the result and has a second model family review it, with Jev making the routing decisions.

mizchi/jev-test-filter

Reads a Git diff, scores every test for whether the change can alter its outcome, and prints the arguments your runner already understands. Every failure path runs the whole suite instead, including a diff that did not fit the state budget.

darkmatter/adhere

Linter for repository rules ordinary linters cannot express: each rule becomes System One questions about the files it covers, run from the CLI or in CI, with Jev as the default provider.

cth9191/jev-compaction-plus

Claude Code compaction that keeps what is still needed word for word and moves the rest to a draft, derived from fast-jev-compaction.

GodsBoy/jev-agent-skill-router

Routes a request to the agent skills it needs, with a confidence floor below which it loads nothing.

0x7067/claude-jev

Claude Code plugin that sends the small judgements — what kind of prompt is this, does it need tools, does this edit break a rule — to a decision model instead of a frontier one.

zjunlp/JevLoop

An agent loop whose seven forks are typed decisions rather than LLM calls, keeping the model for writing. Zero runtime dependencies, and the demo runs offline with no key and no install.

sugarforever/yummy-pi-extensions

Extensions for the Pi coding agent, each released separately, including a Jev-based model router.

racetozero/kiss

Rust terminal coding agent with 44 providers that asks Jev for its own decisions; kiss-coding holds the client.

rhighs/jev-code

Interactive coding CLI and TypeScript SDK built on Jev typed decisions and constrained AST generation. No licence file.

smixs/code-quality

Deterministic quality gate for Claude Code, Codex and pi: Git hooks that block test tampering and secrets, with an optional Jev review that adds findings to the report.

runta-dev/jot

A general-purpose agent loop where Jev picks the next move and Jot runs it. No license file at the time of writing.

transcendr/jevpact

Codex desktop plugin and local service that asks Jev whether the current step has reached a usable conclusion, requests Codex's native compaction at that point, and resumes only the work that is still actionable. macOS development preview.

ckorhonen/jev-lint

Fuzzy linter for coding agents: a hook checks each file the agent writes against your team's rules in .jev-lint/ with Jev and reports violations while the agent is still working.

sedulustech/jint

Semantic linter: rules written in TypeScript build state from Treesitter and ask Jev to classify it, and a callback per rule turns the answers into findings.

francesco0242/matchcn

Semantic index across shadcn-format registries: components are tagged once across six dimensions and committed, and your brief is classified at query time to match against them. Ships as an MCP server.

tyler-dot-earth/patdown

Lints a tree against fuzzy rules kept in one markdown file, behind a provider-neutral interface so the judge can be swapped. Its LICENSE is not a recognised open-source license.

Ice-Hazymoon/jevlint

Semantic lint rules for review questions syntax can't answer, such as a logged secret or an unbounded retry loop, asked of Jev alongside ESLint.

iamtoomas/JevLint

Checks code against conventions written in plain English, in a write-check-fix loop with your coding agent.

jsk4581/jev-blindspot

Side panel for Claude Code and Codex that asks what a prompt leaves unsaid before the agent acts on it.

overbit/fast-jev-compaction-opencode

The fast-jev-compaction approach ported to OpenCode: stale tool history is pruned rather than rewritten by an LLM.

hraness/gobstopper

A local proxy that compacts long Claude Code and Codex sessions as they run, with Jev as one of the scorers that decide what stays; nothing is sent until you choose one.

sedum-dev/sedum

Plain-English browser end-to-end tests on Playwright, cheap enough to run on every pull request: Jev resolves what each sentence refers to and whether a claim holds on the page. TypeSafe, a compatible endpoint, or Cloudflare Clef.

haalp/mergeone

Merge gateway for agent-written pull requests: a decision model, Jev by default, answers typed questions about each PR, and a policy file you own turns the probabilities into pass, hold or block.

joaoh82/yardsort

Desktop app for running coding agents in parallel, each in its own git worktree, with optional Jev judgements on each change. GPL-3.0.

jerryfane/omp-jev-compaction

The fast-jev-compaction approach as a plugin for the omp coding agent: Jev scores every tool call and result, stale ones shrink to a recoverable note, and the rest passes through verbatim.

virolea/lintus

A linter whose rules are sentences in a YAML file: each rule is a question asked of every file it applies to.

nozomi-koborinai/jev-spec

Checks each commit's code against the Markdown specs beside it and fails when they have drifted apart.

Parth1811/JevFlow

Claude Code plugin that has Claude plan work as phases with shell checks; when the agent tries to stop, it runs the checks and asks Jev whether the work is really done. A live viewer follows several agents on one plan.

stratonext/software-factory

Local pipeline that puts a typed judge between stages, including a secret gate, published on PyPI.

frostney/clean-code-review

Reviews every code file in a pull request against a question set drawn from *Clean Code*: Jev answers, Luna writes the review. Runs on eve with Jev through Vercel's AI Gateway; a hosted version is live.

romiluz13/jevmory

Memory for coding agents where every fact is a verbatim quote, graded by Jev into keep, stale, wrong or unsupported, with local SQLite receipts and jevmory audit for existing memory files.

mblode/taste-lint

Design review on every commit: 28 blocking checks and more advisory ones, with the fuzzy ones — vague errors, bare confirm labels, empty states — sent to Jev through Vercel's AI Gateway as review notes.

Nyarlathoteppppp/pi-jev-context

Pi extension that collapses repeated file reads and, opt-in, has Jev mark the passing-test noise in long logs while code keeps every failure; a recall tool returns anything shortened verbatim.

SoyJorgePilo/jev-demo

Three teaching demos where Jev answers and code acts on the probabilities: Jira ticket triage, flagging functions for Claude to review, and routing pull requests to rules, Haiku or Sonnet. Spanish.

jpcamara/judgelint

Linter for conventions written as plain-English bullets in a README: it asks a decision model, Jev by default in its guide, whether each file breaks each bullet and reports violations with line numbers and an exit code.

ismaildasci/claude-referee

Unofficial Claude Code plugin and GitHub Action that sends checkable questions, such as whether a task is really done or which option fits your rules, to Jev and keeps a local log of every check.