The /acmm dashboard assesses any GitHub repository against the AI Codebase Maturity Model (ACMM) — a 5-level framework with 33 feedback loops that measure how well a codebase leverages AI-assisted development.
ACMM was published as an arXiv paper (cs.SE, 2604.09388). The console implementation adds three supplementary sources beyond the core ACMM spec: fullsend, agentic-engineering-framework, and claude-reflect — each contributing additional criteria via a plugin registry.
owner/repo in the Repo Picker at the top (defaults to kubestellar/console).Results are cached for 15 minutes. When no GitHub access is available, demo data kicks in.
Shows the repository’s current maturity level (L1 through L5) as a ring gauge, the matching role description, and a progress bar toward the next level.
| Level | Name | AI Contribution Target |
|---|---|---|
| L1 | Assisted | ~25% |
| L2 | Augmented | ~40% |
| L3 | Collaborative | ~55% |
| L4 | Autonomous | ~70% |
| L5 | Self-Evolving | ~85% |
Weekly bar chart showing the ratio of AI-generated vs human-authored contributions. A level-anchored target slider lets you set a goal (e.g., “aim for L3 = 55% AI”). Click a Use L{n} snap button to jump to that level’s target. The target persists per-repo in localStorage.
Filterable inventory of all 33 ACMM feedback loops (plus criteria from the three supplementary sources), showing which are detected in the repo and which are missing. Filter by source (acmm, fullsend, agentic-engineering-framework, claude-reflect) or by detected/missing status.
Click any missing feedback loop to open a side panel with a description and a Console-specific reference showing how to implement it.
Role-aware top-5 list of missing criteria prioritized for reaching the next level. Each recommendation links to the relevant feedback loop in the inventory card.
The scan runs via a Netlify Function at GET /api/acmm/scan?repo=owner/repo (web/netlify/functions/acmm-scan.mts). It: