Keeps a team's AI coding agents coordinated by capturing engineering decisions and briefing agents before they act, without sharing source code.
Setup
This project documents no single launch command; follow its repository instructions.
Lockstep
lockstep-team-agent
Related
More- RrobrainadelinamartSelf-hosted shared memory for teams and coding agents that records architectural decisions and rejected alternatives and warns before a rejected approach is proposed again.Knowledge & Memory82 Repository stars
- KKrimtokrimto-labsTeam memory for coding agents stored as Markdown in git, with user, team and org levels and hybrid SQLite/sqlite-vec retrieval.Knowledge & Memory7 Repository stars
- SMCP Sequential Thinkingarben-admLocal MCP work notes with persistent sessions, decisions, evidence, risks and resumable tasks.Knowledge & Memory959 Repository stars
- DDecisionNodedecisionnodeRecords development decisions as structured entries and makes them semantically searchable from a local store shared across Claude Code, Cursor and other MCP clients.Knowledge & Memory83 Repository stars
- SSugarRoboticForceOfficialLocal-first persistent memory for AI coding agents, with cross-session context and optional autonomous task execution.Knowledge & Memory98 Repository stars
- RRemembraRemembraOfficialSaves session handoffs built from git facts when an agent stops, so the next coding agent on any machine or MCP client can continue from them.Knowledge & Memory16 Repository stars