microsoft/AI-Engineering-Coach

better agentic engineering

What it solves

It addresses the lack of visibility into how developers use AI coding assistants. Instead of guessing if prompts are effective or if AI is being used efficiently, this tool provides a local dashboard to analyze session logs, identify bad habits (anti-patterns), and measure actual AI-generated code output.

How it works

The tool acts as a read-only analyzer that parses local AI session logs on the user's machine. It processes this data locally to generate metrics, trends, and insights without sending data to external servers. It can be installed as a VS Code extension or run as a canvas within the GitHub Copilot app, utilizing the VS Code built-in Copilot language model API for specific advanced features like skill discovery and context reviews.

Who it’s for

Developers who use AI coding assistants and want to improve their "agentic engineering" skills by tracking their progress, auditing their context management, and refining their prompting techniques.

Highlights

  • Anti-Pattern Detection: Uses 45 rules to flag issues in prompt quality, session hygiene, and tool mastery.
  • Context Health Scoring: Provides agentic readiness checks and audits for instruction files and workspace context maps.
  • Skill Finder: Identifies repeated prompt patterns to help users turn them into reusable skills.
  • Local-First Privacy: All parsing and analytics run entirely on the local machine with no proprietary telemetry.
  • Comprehensive Metrics: Tracks code volume by language, model usage, and activity heatmaps.

Related

  • Dispatch
    GitHub Copilot Internal Architecture: Context Injection and Session Storage Analysis

    A technical deep dive into GitHub Copilot's network traffic and source code reveals how it handles context injection, model routing, and the plaintext storage of user prompts in a local SQLite database.

  • Project
    robertpiosik/CodeWebChat

    CodeWebChat is a privacy‑first VS Code extension that lets you chat with local LLMs about your code. By selecting files, it builds markdown‑based prompts for tasks like editing, searching, generating commit messages, or completing code at the cursor. The model returns structured markdown which the extension can apply directly, all without sending data to external servers.

  • Project
    f/agentlytics

    A local analytics dashboard that unifies conversation history, token costs, and usage patterns from 17 different AI coding editors into one place.

  • Project
    microsoft/agentrc

    A context engineering tool for AI coding agents that automatically generates, measures, and maintains the instruction files and configurations needed for agents to understand a codebase.

  • Project
    kenn-io/agentsview

    A local-first session browser and cost tracker for AI coding agents that centralizes history, search, and analytics across multiple tools.