MCP Quickstart

September 22, 2026 · View on GitHub

Give your AI coding assistant access to 411 indexed failure-recovery lessons from real debugging sessions.

What you get

When your agent encounters an error, it can search MisakaNet for known fixes before wasting time re-debugging:

You: "My pip install keeps timing out on WSL"
Agent: *searches MisakaNet* → finds pip-install-timeout-ssl lesson → applies fix

Copy this to your coding agent

Paste this into Cursor, Claude Code, or Claude Desktop to install and test MisakaNet in one shot:

Please install and test MisakaNet as an MCP failure-memory server.

1. Clone https://github.com/Ikalus1988/MisakaNet
2. Configure it as an MCP stdio server:
   - Command: python3
   - Args: /path/to/MisakaNet/scripts/mcp_server.py
3. Restart your MCP client (Cursor / Claude Code / Claude Desktop)
4. Run the first query: Search MisakaNet for "database locked"
5. Confirm that misakanet_search returns failure-recovery lessons with title, score, and path.

MisakaNet is a failure-memory and recovery layer for coding agents / MCP clients. Local stdio MCP calls work independently of Glama analytics.

Fastest: connect via Glama (no clone, no self-hosting)

The Glama connector proxies to MisakaNet's hosted endpoint — you skip the local setup entirely:

Open the connector page and click Connect through Glama MCP Gateway.
Glama gives you a personal gateway URL: https://glama.ai/endpoints/<your-connection-profile>/mcp.
Add that URL to your client (Claude Code mcp.json, Cursor, VS Code, ChatGPT) as a remote MCP server.

Status on Glama: Ownership verified · Healthy · 7 tools · TDQS A (4.2/5). Prefer local? Use the stdio setup below — it works offline and independently of Glama.

Add to .mcp.json at your project root (project scope — committable, shared with the team), or to ~/.claude.json for user scope. Not .claude/settings.json: that file holds hooks and permissions, and an mcpServers block there is never read (scope table):

{
  "mcpServers": {
    "misakanet": {
      "command": "python3",
      "args": ["/path/to/MisakaNet/scripts/mcp_server.py"]
    }
  }
}

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "misakanet": {
      "command": "python3",
      "args": ["/path/to/MisakaNet/scripts/mcp_server.py"]
    }
  }
}

Cursor

Add to .cursor/mcp.json in your project:

{
  "mcpServers": {
    "misakanet": {
      "command": "python3",
      "args": ["/path/to/MisakaNet/scripts/mcp_server.py"]
    }
  }
}

Docker / GHCR

No local Python setup needed — pull the container and run:

docker pull ghcr.io/ikalus1988/misakanet:2.15.0
docker run -i ghcr.io/ikalus1988/misakanet:2.15.0

Or use in MCP config:

{
  "mcpServers": {
    "misakanet": {
      "command": "docker",
      "args": ["run", "-i", "ghcr.io/ikalus1988/misakanet:2.15.0"]
    }
  }
}

Prerequisites

cd /path/to/MisakaNet
pip install -r requirements.txt

Smoke test

# Verify the server starts
python3 scripts/mcp_server.py --help

# Test search directly
python3 search_knowledge.py "DCO sign-off"

Demo queries

Once connected, try these in your AI assistant:

Query	What you'll find
"Search MisakaNet for DCO sign-off failure"	DCO quickfix workflow
"Find lessons about GitHub token issues"	PAT, credential helper, auth failures
"What does MisakaNet know about pip timeout?"	SSL/proxy timeout fixes
"Search for FANUC error codes"	Industrial robot debugging lessons
"Find feishu API integration issues"	Feishu/Lark API pitfalls

Troubleshooting

Issue	Fix
"No MCP server found"	Check path in config is absolute
"Import error"	Run `pip install -r requirements.txt`
"No results"	Verify `data/lessons.json` exists and is valid
Windows encoding error	Set `PYTHONUTF8=1` or use `python3 -X utf8`