agentleFS
D-SheyD-Shey/autoharness/.agents/skills/kimi-mcp-configuration/SKILL.md

kimi-mcp-configuration

Skill10 starsChanged yesterdayMITView on GitHub

Use when inspecting, editing, validating, or reloading Kimi Code MCP mcp.json configuration, especially when configured or imported servers do not appear in /mcp

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use for missing MCP servers, invalid or misplaced mcp.json files, scope conflicts, and server additions that are not visible in an open …

    line 10
  2. Very specific:

    Use command for stdio and url for HTTP.

    line 17
  3. Very specific:

    Start a new session or restart Kimi after direct file edits.

    line 19
  4. Very specific:

    When six imported servers exceeded the provider's tool-schema size limit, [mcp] default_deferred = true together with [experimental] …

    line 30

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Kimi MCP configuration

When to use

Use for missing MCP servers, invalid or misplaced mcp.json files, scope conflicts, and server additions that are not visible in an open session.

Steps

  1. Resolve the user data root from KIMI_CODE_HOME, falling back to ~/.kimi-code; do not assume the fallback when the variable is set.
  2. Inspect the user-global mcp.json, the project-root .mcp.json, and the current-working-directory .kimi-code/mcp.json as applicable. Record which file declares each server so edits happen at the actual source.
  3. Preserve the top-level mcpServers object and every unrelated entry. A missing file can be initialized with an empty mcpServers object; malformed JSON must be reported rather than overwritten.
  4. Use command for stdio and url for HTTP; add transport: "sse" only for legacy SSE. Keep timeouts as integer milliseconds in the documented range and use environment-variable references instead of new literal secrets.
  5. Show the target path and intended change before writing. Project stdio entries execute at session start, so they belong only in trusted workspaces.
  6. Start a new session or restart Kimi after direct file edits. Servers added mid-session are not registered in that session; use /mcp to inspect connection status.

Reference: https://www.kimi.com/code/docs/en/kimi-code-cli/customization/mcp.html

How to verify

Parse the final JSON and inspect the merged server names by source and precedence. Compare the written entry with the requested command, URL, arguments, environment overlays, and timeouts. In a new session, check /mcp for the server and its connection status; do not treat JSON parsing alone as proof that the server connected.

Facts found here

  • kimi doctor validated config.toml and tui.toml during this task but did not report mcp.json; parse and compare MCP JSON separately.
  • When six imported servers exceeded the provider's tool-schema size limit, [mcp] default_deferred = true together with [experimental] tool-select = true allowed a new session to start and expose the MCP manifests on demand.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.