phemonenon-sys

All Agent Shared Memory — Development skill for Claude Code

Development community

Zero-dependency cross-agent memory + session bridge for Claude Code, Codex, opencode and ZCode: one auditable MEMORY.md injected natively, a stdlib MCP server, history mining into project state files.

How to install All Agent Shared Memory

This entry records only its repository, not the path inside it, so there is no exact command to give. Open phemonenon-sys/all-agent-shared-memory and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What All Agent Shared Memory does

Alternatives in Development

Claude Notifications Go — Cross-platform smart notifications -- 6 types, click-to-focus, context analysis, webhooks 365 ★
Peerbridge MCP — Provider-native adapter and governance layer for auditable multi-agent coding, review, evidence, and release w 136 ★
Tonghuasun Agent — 在 Codex、Claude Code、WorkBuddy、ZCode、OpenClaw 或 DeepSeek Harness 中，直接查询你电脑上的同花顺行情、K 线、持仓、委托和成交数据 110 ★

README

All-Agent Shared Memory

[![ci](https://github.com/phemonenon-sys/all-agent-shared-memory/actions/workflows/ci.yml/badge.svg)](https://github.com/phemonenon-sys/all-agent-shared-memory/actions/workflows/ci.yml) [![license](https://img.shields.io/badge/license-MIT-blue)](LICENSE) [![donate](https://img.shields.io/badge/donate-USDT%20TRC20-green)](https://github.com/phemonenon-sys/all-agent-shared-memory#donate)

![demo](assets/demo.gif)

One small memory file that **every** coding agent reads at session start - Claude Code, Codex, opencode, ZCode, Gemini CLI - plus an append-only log, per-project memory files, and a tiny MCP server. No daemon, no database, no embeddings, no LLM in the memory path. The whole system is a few hundred readable lines: audit it before you trust it with your agents' memory.

every agent session start ─┐
                           ├─ reads  MEMORY.md   (curated hot layer, auto-injected natively)
agent learns a fact ───────┘
                           ├─ writes log\.md        (mem add / MCP memory_add)
                           ├─ writes projects\.md      (mem add -Project name)
                           └─ HOT facts also land in MEMORY.md (-Hot)

Why this one

The agent-memory space is crowded ([deja-vu](https://github.com/vshulcz/deja-vu), [memorix](https://github.com/AVIDS2/memorix), [memmy-agent](https://github.com/MemTensor/memmy-agent), [cass_memory_system](https://github.com/Dicklesworthstone/cass_memory_system), [memsearch](https://github.com/zilliztech/memsearch), [shared-agent-memory](https://github.com/dan-calin/shared-agent-memory), and many more). This one is deliberately the smallest, and it does three things most of them do not:

Zero dependencies, zero binaries. PowerShell and Python-stdlib scripts only. No Go/Rust/Node binary to install, no vector DB, no background service, no API keys. If you can read a script, you can audit your memory layer.
Static injection, not retrieval. Each