[Feature]: Unify the names of the tools

[Makishima]

Problem Description

Hey!

I'm using the context7 MCP server via the Docker MCP Toolkit.

Most MCPs with multiple tools available have a prefix for them.

But context7's tool is registered without one.

Looking at the available tools, I can see several MCP-related tools:

1. MCP_DOCKER_browser_* - Browser automation tools (click, navigate, snapshot, etc.)
2. MCP_DOCKER_code-mode - Create JavaScript-enabled tools combining multiple MCP servers
3. MCP_DOCKER_convert_time - Timezone conversion
4. MCP_DOCKER_fetch - Fetch URLs from internet
5. MCP_DOCKER_get-library-docs - Fetch library documentation
6. MCP_DOCKER_get_current_time - Get current time in timezone
7. MCP_DOCKER_git_* - Git operations (add, checkout, commit, diff, etc.)
8. MCP_DOCKER_mcp-* - MCP server management (find, add, remove, config, etc.)
9. MCP_DOCKER_resolve-library-id - Resolve package names to library IDs
10. MCP_DOCKER_sequentialthinking - For complex problem solving

Proposed Solution

Is it possible to add a "context7_" prefix to make similar tools easier to identify, coming from a single MCP server?

Alternatives Considered

No response

Priority

Nice to have

Additional Context

No response

[fahreddinozcan]

Hey @Makishima, tool names and descriptions are usually not for humans to read, but for LLMs. So we try to be as simple and clear as possible to make it easier for LLM to choose tools properly.

Most MCP clients list the MCP servers' tools on the MCP listing page(like you can see below for cursor), which is meant for humans to debug their MCP client and servers. I'd suggest using that view of your MCP client to understand which tools belongs to which MCP server.

Nevertheless, I will forward this request to team to see what they think how this would affect our tools for LLM usage.

[Makishima]

Thanks for the response! A few points:

1. My question was about CLI LLMs (Claude Code, Codex, Open Code) - not Cursor. There's no "MCP listing page" UI in a terminal. When I'm debugging why a tool wasn't called, I need to read raw tool names in logs/prompts.

2. Docker's Dynamic MCP actually proves the opposite point - with dynamic tool discovery, LLMs use mcp-find to search for servers by name. Clear prefixes like mcp_docker_* help the LLM understand tool provenance and select the right one. If naming is "for LLMs," then consistent prefixes help LLMs distinguish between resolve from context7 vs resolve from another server.

3. When 10+ MCP servers are loaded, seeing context7_resolve_library_id vs just resolve_library_id makes a real difference - both for humans debugging and LLMs disambiguating.