canon-keeper-mcp 0.1.3

MCP server for extracting and persisting learnings from Copilot conversations

Canon Keeper MCP Server

An MCP (Model Context Protocol) server that extracts durable learnings from Copilot conversations and persists them to copilot-instructions.md.

No VS Code extension required! Copilot invokes the MCP tool directly.

Quick Install

# From any project directory
python -m canon_keeper_mcp install

# Or specify a workspace
python -m canon_keeper_mcp install --workspace /path/to/project

The installer will:

1. ✅ Install dependencies (mcp, google-genai)
2. ✅ Configure .vscode/mcp.json with the canon-keeper server
3. ✅ Create/update .github/copilot-instructions.md with:
   • Memory Persistence Protocol (@History directive)
   • Best practices template
   • Session Learnings Log table

Features

• Intelligent Extraction: Uses LLM to identify technical decisions, architectural choices, workarounds, and insights
• Deduplication: Compares against existing learnings to avoid duplicates
• Update Detection: Identifies when a learning updates/supersedes an existing entry
• Markdown Formatting: Returns properly formatted table rows for the Session Learnings Log

Usage

After installation, reload VS Code and use these trigger phrases in any Copilot chat:

User: @History save what we learned
User: save this learning  
User: remember this
User: add to memory

Copilot will:

1. Gather the full conversation history
2. Call the MCP tool to extract learnings
3. Deduplicate against existing entries
4. Append new learnings to your copilot-instructions.md
5. Report what was saved/skipped

Manual Installation

If you prefer to install manually:

1. Install dependencies

pip install mcp google-genai

2. Configure MCP server

Add to .vscode/mcp.json:

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "mcpServers": {
    "canon-keeper": {
      "command": "${workspaceFolder}/.venv/Scripts/python.exe",
      "args": ["-m", "canon_keeper_mcp"],
      "cwd": "${workspaceFolder}"
    }
  }
}

3. Add directive to copilot-instructions.md

Add this to your .github/copilot-instructions.md:

### Memory Persistence Protocol (@History)

**Rule:** When the user includes `@History`, `save this`, `remember this`, or `add to memory` in any message:

1. Read the current `copilot-instructions.md` file content
2. Collect the full conversation history from this session
3. Call MCP tool `canon_keeper.extract_and_save_learnings` with:
   - `conversation`: The complete conversation text
   - `current_instructions`: The full content of copilot-instructions.md
4. If `markdown_to_append` is non-empty, append it to the Session Learnings Log table
5. Report what was saved and what was skipped as duplicates

**Trigger phrases:** `@History`, `save this`, `remember this`, `add to memory`

4. Add Session Learnings Log table

## Session Learnings Log

| Date | Topic | Decision | Rationale |
|------|-------|----------|----------|

MCP Tools

extract_and_save_learnings

Extracts learnings from conversation and deduplicates against existing entries.

Input:

• conversation (string): Full conversation text
• current_instructions (string): Current copilot-instructions.md content

Output:

{
  "status": "success",
  "message": "Found 2 new learning(s), skipped 1 duplicate(s).",
  "new_learnings": [...],
  "duplicates_skipped": [...],
  "markdown_to_append": "| 2026-01-17 | Topic | Decision | Rationale |",
  "target_section": "Session Learnings Log"
}

check_learning_exists

Check if a specific learning already exists.

Input:

• topic (string): Learning topic
• decision (string): The decision/learning
• current_instructions (string): Current copilot-instructions.md content

Output:

{
  "exists": true,
  "similar_to": "Existing Topic Name",
  "reason": "Semantically equivalent to existing entry"
}

LLM Providers

Supports:

• Google GenAI (default): Uses gemini-2.0-flash-001
• OpenAI (fallback): Uses gpt-4o-mini

Set appropriate API keys:

• GOOGLE_API_KEY for Google GenAI
• OPENAI_API_KEY for OpenAI