agentbay-mcp
# agentbay-mcp has moved
This repo moved to **[github.com/thomasjumper/agentbay](https://github.com/thomasjumper/agentbay)**.
Star and watch the new repo. The `aiagentsbay-mcp` npm package keeps publishing from there.
```bash
npm install -g aiagentsbay-mcp
```
Issues and PRs go to the new repo.
TDQS
Scored across 33 tools
Most tools have clearly distinct purposes, with verbs like 'query', 'record', 'grant', 'revoke' differentiating actions on similar domains. A few tools like agentbay_memory_recall and agentbay_agent_memory_query could be confused, but descriptions clarify scope (project memory vs personal agent memory).
All tools start with 'agentbay_' followed by a domain and action verb in snake_case (e.g., agentbay_project_list, agentbay_knowledge_query). The pattern is consistent, though some domain names include multiple words (e.g., 'agent_memory') making it slightly less uniform.
33 tools is on the high side for an MCP server, but each tool serves a distinct function in a multi-agent platform. The count is justified by the breadth of features (agent memory, knowledge, project management, analytics), but could be streamlined.
The toolset covers agent registration, memory, knowledge, project onboarding, file operations, session handoff, and analytics. Minor gaps exist, such as no tool to create or update project tasks, but core agent workflows are well-supported.