Cross-agent coordination for AI coding agents. One daemon, one database — so multiple agents in one repo can see each other's plan, tasks, and file claims.
Shared plan, tasks, and file claims for AI coding agents working on one project.
Three Claude Code sessions, an OpenCode window, and a Factory Droid run can all be editing the same repository right now — and none of them knows the others exist. They duplicate work, overwrite each other, and lose the plan every time a session ends.
succubus is the shared brain they all read and write: one daemon, one database, and a dashboard that shows who is doing what in real time.
- Names every agent. Each session adopts an identity — ORION, VESPER, KESTREL — and keeps it across restarts and context compaction. Three concurrent sessions of the same tool get three distinct names; a resumed session gets its own name back. - Holds the plan and the task board. Agents read the active plan on every turn and claim tasks so two of them never start the same work. - Leases files. An agent claims the paths it is about to edit. Another agent asking for the same path is told who holds it. Leases expire on their own, so a crashed session never blocks a file permanently. - Gives agents somewhere to ask. An agent that is unsure posts a question instead of guessing. Other agents see it on their next turn; you see it immediately. - Refuses to be ignored. Coordination state is injected into each agent's context on session start and on every prompt, so an agent cannot claim it did not know — and it is told to write back, not only to read. - Never blocks your work. If the daemon is down, hooks exit silently and MCP tools return advisory text. succubus degrades to nothing rather than breaking your session.
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/enowdev--succubus.html)
Render agent-written Markdown as a live local page, and send the human's click back to the agent as typed data. Human-in-the-loop approval gates for coding agents.
sainitish1609/mcp-guard🛡️ Ultra-fast local security firewall, secret sanitizer, and context token compressor for Claude Code, Cursor, and MCP AI agents.
luckeyfaraday/frontier-orchestratorMulti-agent orchestration for Claude Code: MCP server + skill that keeps Claude as lead engineer while delegating backend work to OpenAI Codex and frontend/design work to Kimi
valtors/relayAll-in-one MCP server. Agents get memory, tools, and coordination in one binary.
oyekamal/claude-code-flow-visualizer🔍 Visualize any Claude Code agent harness — CLAUDE.md, subagents, skills, commands, hooks & MCP servers — as an interactive flow graph. Import from GitHub, ZIP, or local folder. 100% client-side.
replenhq/replenSmarter AI Development Workflows
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.