What is MOMO CODE and what does it do?

An AI coding agent that evolves with you, built on opencode, with a dual-speed self-evolution system based on Pioneer Agent. Its CLI runs interactive REPL sessions, one-shot tasks, and JSONL output.

How does MOMO CODE use Jev?

Its v2.0.0 release adds a JEV decision layer: five decision gates with shadow mode and ECE calibration. The README section describing the layer is cut off in the digest, so the exact questions and endpoints are unknown.

How do I start the MOMO CODE CLI?

Run momo for the interactive REPL, momo followed by a task string for one-shot mode, momo exec for explicit single execution, or momo --json for structured JSONL output. Built-in commands include /help, /exit, and /cot.

What is the DecisionPort contract in MOMO CODE?

A type contract whose decisions are choice, bool, score, or abstain, paired with TaskResult and EventEnvelope types for structured logging. The README excerpt does not show its full shape or how it maps to Jev answers.

What is the RSI self-evolution loop?

A three-level (R1/R2/R3) self-evolution loop with budget guardrails and STOP protection, evaluated by momo-bench with three layers — selection, holdout, and canary — plus seven adversarial tests.

Is MOMO CODE stable enough to use?

The README marks v2.0.0 as experimental and under active development, with the major architecture upgrade dated 2026-10-11. Treat the JEV layer and RSI loop as moving targets.

Who maintains MOMO CODE and how popular is it?

momozi1996 publishes the TypeScript repo; the digest shows 1,167 stars, 7 forks, creation on 2026-06-16, and a push on 2026-10-17. The license field is NOASSERTION, so no standard license was detected.

Jev directory

CommunityJev Agents & MCP

MOMO CODE

An experimental coding agent CLI built on opencode whose v2.0.0 architecture adds a JEV decision layer with five decision gates, shadow mode, and ECE calibration, plus a three-level RSI self-evolution loop and six Pi-framework extensions.

See highlights Open github.com

Stars: 1.2k
Language: TypeScript
Reading time: 2 min
Published by: Community
Author: momozi1996
Last verified: Oct 2026

momozi1996/momo-code

User repository on GitHub

View on GitHub

MOMO CODE — RSI *JEV*Pi --AI coding agent that evolves with you

Forks: 7
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

Tagscommunity github typescript coding-agent cli

Works forJev for Code Jev for Decision gating

Used inJev for Software & DevTools

Languages: TypeScript
Platforms: CLI
Capabilities: Open source

Highlights

v2.0.0 (2026-10-11) adds the JEV decision layer: 5 decision gates with shadow mode and ECE calibration.
Integrates @earendil-works/[email protected] as an extension dependency; the README stresses it is not a fork.
DecisionPort type contract covers choice/bool/score/abstain; TaskResult and EventEnvelope types shape structured logs.
RSI self-evolution loop has three levels (R1/R2/R3) with budget guardrails and STOP protection; momo-bench adds 7 adversarial tests.
CLI: momo opens a REPL, momo exec runs one task, momo --json emits JSONL, and Ctrl+T toggles the chain-of-thought panel.

Quickstart

bash

momo
momo "帮我写一个快速排序"
momo --json "任务描述"

Watch out

README reviewed 2026-10-17 and truncated in the middle of the JEV decision layer section, so calibration numbers and the layer's internals are not in the digest. The project labels itself experimental, with v2.0.0 in a major architecture upgrade, and no install command appears in the excerpt.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

X: Decision outputs in Grep

Hacker News: jevmem keeps Claude Code project memory by asking Jev which turns are worth saving.

avinashjetwani61 pointsJevmem – automatic project memory for Claude Code, built on JevRead the thread on Hacker News

X: A usage router for Grok Bot

X: The compaction plugin, in Chinese

From the Jev guides

Original write-ups on what this entry is for.

MCP tool routing: Jev in front of the tool call

Oct 16, 2026

Tool menus can outgrow the context that reads them. The catalog's Jev-based routers and gates narrow the list: benchmarked tool search, per-tool scoring, fail-closed approvals, and WebMCP in the browser.

Jev Agents & MCP#mcp#tool-routing#agents

Read article

Approval policies for coding agents: hooks, guards, and gateways

Oct 16, 2026

How the Jev guard family turns coding-agent permissions into auto-approve, deny, or ask: hooks that enforce, guards that fail safe, and gateways that gate generation.

Jev Agents & MCP#coding-agents#guardrails#hooks