AI-Engineering-Coach

Why we included this project

Most teams adopting AI coding assistants cannot say whether that time is actually making their engineers better. AI Engineer Coach answers that with a local dashboard that reads session logs from the assistants you already use, Copilot, Claude, Codex, and the Copilot CLI among them, then turns the raw activity into practice scores, weekly trends, per-language and per-model code counts, and activity heatmaps. A rule engine checks your sessions against 45 anti-patterns in prompt quality, session hygiene, code review, and context management, and it flags repeated prompts worth saving as reusable skills. Because all parsing and analytics happen on your machine, it works fully offline for developers and team leads who want honest self-review without sending telemetry anywhere. If you are adopting agentic coding and want more than gut feeling to judge your progress, this gives you a concrete view of what is improving and what is not.