auggie-mcp

Retrieve semantic codebase leads and recover external task indexes via Auggie MCP.

Updated Apr 15, 2026
One-click install
npx skills add https://github.com/AoManoh/project_templates --skill auggie-mcp
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: auggie-mcp
Source: https://github.com/AoManoh/project_templates/tree/main/skills/auggie-mcp
Command: npx skills add https://github.com/AoManoh/project_templates --skill auggie-mcp

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents AI from making unsupported claims by using Auggie/Augment MCP to find relevant source evidence and to restore external task progress as leads rather than facts.

Core Features & Use Cases

  • Semantic codebase retrieval: Find relevant files, call chains, and impact/risk leads across modules.
  • Source-grounded verification: Forces confirmation by opening real files before editing or concluding.
  • External task index recovery: Restores candidate next steps and continuity when context is lost, without treating it as truth for requirements or implementation details.
  • Degrade gracefully: Falls back to file listing, precise text search, and real file reading when MCP is unavailable or results are too broad.

Quick Start

Ask your AI to run the auggie-mcp skill to locate where a specific behavior is implemented, then verify the findings by opening the exact files and line ranges before proposing any changes.

Frequently Asked Questions about auggie-mcp

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I find relevant code and call chains across modules during impact analysis?

Semantic codebase retrieval finds relevant files, call chains, and impact or risk leads across modules. You must verify these findings by opening the exact source files before proposing edits or conclusions.

Why does my AI make unsupported claims about codebase behavior during refactoring?

Source-grounded verification prevents unsupported claims by forcing the AI to open real files and confirm line ranges before editing. This ensures conclusions are based on actual source evidence rather than assumptions.

How do I recover task progress and next steps when context is lost in a debugging session?

External task index recovery restores candidate next steps and continuity when context is lost. Treat the recovered task state as leads to verify rather than absolute truth for requirements or implementation details.

What should I do if semantic code retrieval returns overly broad or unavailable results?

Graceful degradation falls back to file listing, precise text search, and real file reading when semantic retrieval is unavailable or results are too broad. This enforces strict downgrade paths to maintain workflow continuity.

Can I use semantic search to locate code without confirming the project root directory?

No, root-directory confirmation is required before initiating semantic codebase retrieval. This ensures the search scope is correctly bounded and the retrieved call chains or file references are accurate.

Does source-grounded verification work for cross-module impact analysis and code review?

Yes, source-grounded verification applies to development, debugging, refactoring, and review scenarios where code location, call chains, or historical task states are uncertain. It forces real file verification before concluding.