Advisory Engine & Maturity Model
Evaluates your environment against 11 categories producing a maturity score (0–67 points).
Categories
| Category | What's Evaluated | Max Points |
|---|---|---|
| Policy Hygiene | Policy files, deny rules, deprecated fields, v0.40+ mode coverage, YOLO safeguards, MCP governance, denyMessage, headless rules | 15 |
| MCP Governance | Server allowlists, per-server governance policies | 15 |
| GEMINI.md Quality | Global context file exists, length, structured sections | 15 |
| Skills Optimization | Skill count, diversity, extraction state health | 15 |
| Settings Hardening | Model pinning, experimental flags, sandbox config | 15 |
| Hooks | Pre/post hooks for tool calls | 10 |
| Extensions | Extension count, catalog diversity | 15 |
| Context Architecture | .geminiignore, repo-level configs, context layering | 15 |
| Memory Architecture | 4-tier memory hierarchy coverage, bloat, cross-duplication, autoMemory | — |
| Skill Extraction | Inbox backlog, extraction activity, stale locks, pending patches | — |
| Admin Policies | System-level policy accessibility | — |
Memory, Extraction, and Admin categories produce advisory recommendations but don't contribute to the numeric score.
Maturity Tiers
| Tier | Score | Description |
|---|---|---|
| 🌱 Getting Started | 0–33 | Basic installation |
| 🔧 Intermediate | 34–46 | Active usage, some governance |
| ⚡ Advanced | 47–60 | Strong policies, skills, MCP governance |
| 🏆 Expert | 61–67 | Full-stack with hooks, extensions, context architecture |
v3.5 Advisory Checks
Policy Hygiene (v0.40+)
| Check | Severity | Trigger |
|---|---|---|
| No approval mode coverage | warning | Rules exist but none specify modes |
| YOLO without safeguards | critical | YOLO mode rules found with no deny rules |
| Plan mode unprotected | warning | No rules scoped to plan mode |
| Ungoverned MCP servers | warning | MCP servers in settings with no policy rules scoping mcpName |
| Missing denyMessage | info | Deny rules without custom explanation text |
| No headless-mode rules | info | No rules use interactive = false |
| commandRegex usage | info | Regex patterns in rules (brittleness warning) |
| Admin policies inaccessible | info | System policy dirs exist but can't be read |
Memory Architecture
| Check | Severity | Trigger |
|---|---|---|
| No global memory | warning | ~/.gemini/GEMINI.md missing |
| Memory bloat | warning | Any MEMORY.md > 5,000 words |
| Unused tier | info | Only 1 of 4 tiers in use |
| Cross-tier duplication | info | Same bullet in global and project memory |
| autoMemory disabled | info | settings.json has autoMemory: false or absent |
| Single-tier usage | info | Only using Tier 1 (global) with no project memory |
Skill Extraction
| Check | Severity | Trigger |
|---|---|---|
| Inbox backlog | warning | ≥3 skills awaiting approval |
| Extraction inactive | info | Last run > 14 days ago |
| Never run | info | No extraction-state.json found |
| Stale lock | warning | .extraction.lock > 30 minutes old |
| Pending patches | info | Patch files awaiting application |
Output Formats
- TUI: Maturity Dashboard (scores, breakdown, recommendations)
- Report: Advisory Recommendations section in markdown
- JSON:
maturityobject ingemini-env-manifest.jsonwithscore,tier,breakdown,recommendations