User repository on GitHub
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Agents & MCP
Jev Flash Review
Rule-based code review for AI coding agents, powered by TypeSafe Jev. Three skills — review a PR, review anything, or loop until the engine reports clean — share one local MCP review engine that never reads your repository; the calling agent supplies the diff and the business context.
- Stars
- 1
- Language
- TypeScript
- Reading time
- 3 min
- Published by
- Community
- Author
- thebous
- Last verified
- Oct 2026
Used inJev for Software & DevTools
- Languages
- TypeScript
- Capabilities
- Open sourceRuns locallyRequires API key
Highlights
- Three skills — review-pr, review-free and review-loop — share one review_diff MCP tool; if it is missing they stop rather than substitute a manual review.
- Every rule in src/rules/ becomes a typed YES, NO or N/A Choice question, batched per chunk and run in parallel; a rule keeps its most severe outcome.
- Locations with top confidence under 0.55 are dropped; survivors pass a confirm question and earn a none, minor, significant or critical impact rating.
- Setup needs Node.js 24+, a TypeSafe API key in a gitignored .env, and one npm run bundle because dist/ is not committed.
- Claude Code installs from the lucvalse/jev-flash-review marketplace; Codex registers the engine manually in ~/.codex/config.toml.
Quickstart
bash
npm install
echo 'TYPESAFE_API_KEY=...' > .env
npm run bundleWatch out
Copy reviewed against the repo README on 2026-10-15; the README publishes no accuracy or benchmark numbers and no license is listed in the GitHub metadata. Note that the README's Claude Code install commands point at a lucvalse/jev-flash-review marketplace entry while the manifest repo is thebous/jev-flash-review.
Questions about this entry
Short answers drawn from the linked project or page.
Rule-based code review for AI coding agents, powered by TypeSafe Jev. Three skills — review-pr, review-free and review-loop — ship to your agent and share one local MCP review engine that never reads your repository.
Each rule in src/rules/ becomes a typed Choice question with YES, NO or N/A options and an applies_if condition; questions are batched per chunk and rule set and run in parallel, and a rule takes its most severe outcome across chunks.
No. It is local-first: the engine never reads your repository. The calling agent curates the input — a unified diff plus taskContext describing the task's business purpose, boundaries and invariants — and the engine returns structured verdicts over what it is given.
Run /plugin marketplace add lucvalse/jev-flash-review and /plugin install jev-flash-review@jev-flash-review. The skills appear as jev-flash-review:review-pr, :review-free and :review-loop, and the bundled .mcp.json registers the review engine automatically.
Node.js 24+, a TypeSafe API key from console.typesafe.ai, and an authenticated gh for the PR skill. After cloning, run npm install, write TYPESAFE_API_KEY into .env, and run npm run bundle once so dist/server.js exists for the MCP server.
One JSON object: results holds the full matrix for every rule including dropped ones, violations holds confirmed findings only, and each outcome carries answer, probability, confidence, severity, question text, evidence locations and impact, with a summary of total, yes, no, n/a, blockers and dropped.
It does not silently fall back to a manual review: if the review_diff tool is unavailable the skills stop and tell you how to register the engine. The README publishes no benchmark numbers, and dist/ is intentionally not committed.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: jev-review: a local-first score loop for coding agents
X: Jev in front of the HeyGen MCP
Hacker News: Thread on Selmar's Jev code-review calibration notes and ordering.py comparisons.
Lectem2 pointsCalibrating Jev as a Code ReviewerRead the thread on Hacker NewsX: Playwright CLI + Jev vs Playwright MCP
More Jev resources like this
Six evidence-oriented tools for the pi and omp coding agents, also usable from any MCP client: navigate unfamiliar code, ask typed questions about repository evidence, review completed changes, and select existing tests.
Jev Agents & MCP#community#catalog-import#logicrw
Community
nomenak/jev-toolsTypeScript
nomenak/jev-toolsTypeScriptAn MCP server that gives coding agents only the JS/TS code a task needs: the TypeScript compiler follows calls, types, callers and tests, Jev judges relevance, and results come back within a token budget.
Jev Agents & MCP#community#catalog-import#logicrw
Community
123wwwa/jevtraceTypeScript
123wwwa/jevtraceTypeScriptA programmable runtime that gives the Pi coding agent one fabric_exec tool: Pi writes a checked TypeScript program that composes core tools, MCP servers and agents into branches and fan-out, executed in an isolated QuickJS sandbox with approvals, budgets and durable coordination.
Jev Agents & MCP#community#catalog-import#logicrw
Community
monotykamary/pi-fabricTypeScript
monotykamary/pi-fabricTypeScriptA very WIP tool for jev-ification from Harry Solovay: lessons from coding sessions become approved Markdown traits, and a configurable decision model evaluates code one trait at a time while an MCP server handles group-aware approval.
Jev Agents & MCP#community#github#typescript
Community
harrysolovay/traits-pocTypeScript
harrysolovay/traits-pocTypeScriptOpen-source Codex plugin that consults TypeSafe Jev about tools, models, tasks, skills, context, and strategies, diagnoses failed commands, and checks completion claims against evidence. The current package is version 0.4.0.
Jev Agents & MCP#community#catalog-import#logicrw
Lets AI agents use Windows applications through MCP or the command line: hand off a bounded task and Jev observes windows, chooses controls, types supplied text and navigates locally, returning a result or a request for help.
Jev Agents & MCP#community#catalog-import#logicrw
Community
jacks3tr/jev-desktopPython
jacks3tr/jev-desktopPythonFrom 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.