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

claude-kimi-mcp-normalization

Skill10 starsChanged yesterdayMITView on GitHub

Use when importing or normalizing Claude Code MCP server declarations into Kimi Code without overwriting existing entries, especially after an import appears to succeed but servers are absent from Kimi

  • Reads credentials

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use for Claude Code MCP imports, source-to-target scope decisions, field conversion, disabled source servers, and unresolved environment …

    line 10
  2. Distinctive:

    Keep user declarations user-global and project-local declarations project-local.

    line 16

≈ “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

Claude-to-Kimi MCP normalization

When to use

Use for Claude Code MCP imports, source-to-target scope decisions, field conversion, disabled source servers, and unresolved environment placeholders.

Steps

  1. Resolve CLAUDE_CONFIG_DIR before choosing the Claude home. Read Claude's user configuration only for an MCP import and inspect both user-scope mcpServers and the resolved project-root or current-directory entry under projects.
  2. Skip source-project names listed in disabledMcpServers; do not silently re-enable them or convert that project choice into a global disabled entry.
  3. Keep user declarations user-global and project-local declarations project-local. Preserve every existing Kimi entry and keep the existing Kimi entry on a name collision.
  4. Map stdio, HTTP, and SSE to Kimi's mcpServers shape. Map timeout to toolTimeoutMs and alwaysLoad: true to deferred: false; omit inferred transport except explicit SSE.
  5. Drop Claude-only policy fields rather than pretending they work in Kimi. Report unsupported transports and entries blocked by unresolved placeholders.
  6. Convert an Authorization: Bearer ${VAR} header to bearerTokenEnvVar: "VAR". Drop pass-through environment declarations because Kimi inherits the parent environment; never substitute a secret's current value into the preview or target.
  7. Warn about apparent literal credentials and trusted-workspace requirements for project stdio servers. Direct file changes require a new session before /mcp can show them.

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

How to verify

Compare source and normalized server names, scopes, launch commands or URLs, arguments, environment overlays, and timeouts. Confirm every skipped entry has a reason and every pre-existing Kimi entry is unchanged. Parse the final target JSON, then check /mcp in a new session for actual loading and connection status.

Facts found here

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.