# AI Engineering Coach

> Updated 2026-09-12 · type: tool · category: mcp-agents · status: active · rev 1

AI Engineering Coach reads your local agent session logs and scores how you actually work with coding agents, across 45 anti-pattern rules, with zero telemetry.

- Open source: yes (MIT)
- Self-hostable: yes
- Pricing model: free
- Best for: Anyone spending real money on coding agents who wants evidence about their own usage patterns before buying more seats or tokens.
- Last verified: 2026-09-12

- **Canonical:** https://gtmstacker.com/registry/tool/ai-engineering-coach/
- **Source:** [github · microsoft/AI-Engineering-Coach](https://github.com/microsoft/AI-Engineering-Coach)
- **Tags:** mcp-agents, agent-ops, developer-analytics, vscode, local-first
- **Repository:** https://github.com/microsoft/AI-Engineering-Coach

## Is AI Engineering Coach open source?

Yes, AI Engineering Coach is open source under the MIT license.

## How much does AI Engineering Coach cost?

AI Engineering Coach is free to use.

## Can I self-host AI Engineering Coach?

Yes, AI Engineering Coach can be self-hosted (the source is available under the MIT license).


---

Open-source VS Code extension from Microsoft (MIT; also runs in Cursor and Antigravity) that reads your LOCAL AI-coding session logs — GitHub Copilot, Claude Code, Codex CLI, OpenCode, Xcode — and scores your workflow across prompt quality, session hygiene, code review, tool mastery and context management, with 45 anti-pattern detection rules. Read-only, zero telemetry.

## Provenance

- MIT + official Microsoft org independently WebFetch-verified 2026-09-12. Open-sourced May 2026; surfaced via the 2026-09-12 viral-posts brief (macro-tier resurgence) — cataloged for utility, with its actual release date stated plainly.
- Curated from the GTM Stacker signal registry (2026-09-12 pass: daily pull + viral-posts brief); license independently WebFetch-verified 2026-09-12.

## Why it matters for a GTM stack

Agent spend audits (the theme of this week's Spotify shunt entry) need usage data, and this is the local, no-telemetry way to get it: which sessions sprawl, which prompts rework, where context gets wasted. Measure first, then route or ration.