What this error means

MCP server launches correctly and sends valid initialize response over stdin/stdout, but Claude connection closes immediately afterward with logs showing server started and connected then disconnect is a Claude Code failure pattern reported for developers trying to debug why custom mcp server built in c#/.net exits immediately after successful initialization when connecting to claude desktop — find proper json-rpc 2.0 lifecycle handling. Based on the imported evidence, treat this as a tool-specific troubleshooting page rather than a generic API error.

Why this happens

SO Q79646317 (score 2). Custom MCP server in C# .NET 9 builds correctly, sends valid JSON-RPC 2.0 initialize response, but Claude Desktop terminates the connection immediately. Distinct from -32001 timeout (Q79703292) — here the error happens within seconds rather than 60s. No good solution found yet, suggesting content gap potential. Category mapping: Claude Code → AI Coding Tools per SKILL.md rules.

Common causes

  • SO Q79646317 (score 2). Custom MCP server in C# .NET 9 builds correctly, sends valid JSON-RPC 2.0 initialize response, but Claude Desktop terminates the connection immediately. Distinct from -32001 timeout (Q79703292) — here the error happens within seconds rather than 60s. No good solution found yet, suggesting content gap potential. Category mapping: Claude Code → AI Coding Tools per SKILL.md rules.

Quick fixes

  1. Confirm the exact error signature matches MCP server launches correctly and sends valid initialize response over stdin/stdout, but Claude connection closes immediately afterward with logs showing server started and connected then disconnect.
  2. Check the Claude Code account, local tool state, and provider configuration involved in the failing workflow.
  3. Compare the failing environment with a known working setup, then change one configuration value at a time.

Platform/tool-specific checks

  • Verify the command, editor, extension, or API client that produced the error.
  • Compare local settings with CI, deployment, or editor-level settings when the error appears in only one environment.
  • Avoid deleting credentials, local model data, or project settings until the failing scope is clear.

Step-by-step troubleshooting

  1. Capture the exact error message and the command, editor action, or request that triggered it.
  2. Check whether the failure is account/auth, quota/rate, model/provider, local runtime, or deployment configuration.
  3. Review the source evidence below and compare it with your environment.
  4. Apply one change at a time and rerun the smallest failing action.
  5. Keep the working fix documented for the team or deployment environment.

How to prevent it

  • Keep provider/tool configuration documented.
  • Record non-secret diagnostics such as tool version, provider name, model name, and command path.
  • Add a lightweight check before CI or production workflows depend on the tool.