sma-framework

Sma

Name: Sma
Author: sma-framework

AI community

SOURCE https://github.com/sma-framework/sma

Description

Memory and accountability for AI coding agents. Layered memory that arrives at the moment of action, multi-terminal coordination without a server, and a trust spine where every "done" is settled by a re-runnable check. Local-first: plain files + git, zero LLM in the hot path. Works with Claude Code.

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

SMA — Shared Memory & Automation

**Your AI coding agent forgets, overstates, and does not scale into a team. SMA is the local-first control plane that fixes all three at once: layered memory that arrives at the exact moment of action, multi-terminal coordination without a server, and a trust spine in which every "done" is settled by a script, re-derived by a blind verifier, and blocked from shipping when it is false.**

[Русская версия → README.ru.md](README.ru.md)

🗺️ [Open the live system map →](https://sma-framework.github.io/sma/master-graph.html)

Every subsystem of SMA on one interactive page — the fastest way to see how everything connects.

🆕 [What's new in v5.7.0 — the design stage →](#whats-new-in-v570--the-design-stage)

The phase graph gains a fifth stage between the plan and the code: the thing is drawn, a short contract is written beside the drawing, and **execution does not start until a person confirms it**. With it come two new rol