Browse / Software development / conn2flow-ai-workspace
conn2flow-ai-workspace
otavioserra
Autonomous Multi-Agent Memory-Driven Development (MDD) Framework.
Open otavioserra/conn2flow-ai-workspace → Checked against the four rules.
agentic-aiai-codingantigravityclaude-codecodexdeveloper-toolsmddmemory-driven-developmentmulti-agent-systemsvscode-extension
Works with any agent
ships AGENTS.md, the cross-tool instruction file.
In its own words
Triad Orchestration (Architect, Executor, Reviewer), 44 Canonical AI Skills, Python CLI/Hub & VS Code Dev Tools.
The Memory-Driven Development operating system for multi-agent engineering
Português · Documentation
Long-running AI projects lose context between sessions. Memory-Driven Development (MDD) keeps specifications, decisions, execution evidence and reusable procedures in Git so the next agent can resume from a small index and verified sources.
This workspace provides the foundation for an Architect–Executor–Reviewer triad, 45 canonical skills, a VS Code panel and an MCP bridge. Its Python roadmap adds the MDD Client CLI/Daemon for local memory and the MDD Hub for reports and documentation monitoring. The memory approach can serve any project; Conn2Flow-specific skills support its PHP and resource ecosystem.
How it fits together
Four layers separate episodes, approved knowledge, procedures and raw observations. Hierarchical indexes guide retrieval; dual archives preserve both a concise summary and the original evidence.
Solid connections describe the current workflow; dotted connections show the Python components approved for development. At this checkout, their packages are not yet present. See the Python guide for availability.
The repository's own description, then 38% of its README. Read the rest on GitHub ↗
Read from the repository's git tree on 2026-08-17.