odoo-analyzer
AI-powered Odoo engineering platform providing static analysis, domain knowledge, and 14 MCP tools for code review, model exploration, and security auditing of Odoo modules.
README
Odoo Best Practices — AI-powered Odoo Engineering Platform
Static analysis engine + domain knowledge platform + MCP server for Odoo module development. Covers 136 rules, 12 anti-patterns, 6 Odoo versions, and provides 14 MCP tools for AI-assisted code review.
Quick Start (5 minutes)
<details> <summary><b>🐍 Python (recommended for full features)</b></summary>
# 1. Clone
git clone https://github.com/FoxPink-dev/odoo-best-practices.git
cd odoo-best-practices
# 2. Run analysis on your Odoo addon (use python3 on Linux, full path on Windows)
python -m analyzer.cli /path/to/your/addon --check
</details>
<details> <summary><b>📦 npm (quick start — config generation)</b></summary>
npx @foxpink-dev/odoo-best-practices init /path/to/your/addon
</details>
Expected output:
Models: 12
Fields: 142
Violations: 3 violations found
CRITICAL: 1
HIGH: 1
MEDIUM: 1
LOW: 0
If you see output like above, the analyzer is working. You just ran your first Odoo analysis in 5 minutes.
Try with demo data
python -m analyzer.cli tests/fixtures/demo_addon --check --format json
GitHub Action Setup
1. Add workflow file
Create .github/workflows/odoo-review.yml:
name: Odoo Review
on:
pull_request:
paths:
- '**.py'
- '**/__manifest__.py'
- '**/*.xml'
- '**/security/*.csv'
jobs:
analyze:
runs-on: ubuntu-latest
permissions:
contents: read
pull-requests: write
security-events: write
steps:
- uses: actions/checkout@v4
- name: Odoo Static Analysis
uses: ./.github/actions/odoo-review
with:
addon-path: '.'
fail-on-critical: 'true'
fail-on-high: 'true'
baseline: 'true'
github-token: ${{ secrets.GITHUB_TOKEN }}
- name: Upload SARIF to Code Scanning
uses: github/codeql-action/upload-sarif@v3
with:
sarif_file: odoo-review-results.sarif
category: odoo-analyzer
continue-on-error: true
2. PR Results
After running, you will see in the PR:
| Feature | Display |
|---|---|
| PR Comment | Violations table grouped by severity with file:line |
| Annotations | Inline error/warning markers on each code line |
| Code Scanning | SARIF upload → GitHub Security tab |
┌─────────────────────────────────────────────┐
│ ## Odoo Analysis Report │
│ | Severity | Count | │
│ |----------|-------| │
│ | 🔴 CRITICAL | 1 | │
│ | 🟠 HIGH | 1 | │
│ | 🟡 MEDIUM | 1 | │
│ | **Total** | **3** | │
│ │
│ ### 🔴 CRITICAL (1) │
│ - **`models/sale.py:42`** — search() inside │
│ loop │
│ <sub>rule: `search-inside-loop`</sub> │
└─────────────────────────────────────────────┘
Tip: On first run against a legacy repo, use
generate-baseline: 'true'to baseline existing violations.Note: PyPI publish is currently skipped (npm-only distribution). See
publish-npm.ymlfor npm CI publish workflow.
MCP Server (AI Agent Integration)
Cursor
{
"mcpServers": {
"odoo-analyzer": {
"command": "python",
"args": ["-m", "analyzer.mcp_server", "/path/to/addon"]
}
}
}
Claude Code
{
"mcpServers": {
"odoo-analyzer": {
"command": "python",
"args": ["-m", "analyzer.mcp_server", "/path/to/addon"]
}
}
}
OpenCode
{
"mcpServers": {
"odoo-analyzer": {
"command": "python",
"args": ["-m", "analyzer.mcp_server", "/path/to/addon"]
}
}
}
14 MCP Tools Available
| Tool | Description |
|---|---|
analyze_module |
Full module analysis: models, views, security, dependencies |
search_model |
Model definition with fields, methods, inheritance |
search_view |
View definition with inheritance chain |
search_action |
Action definition with linked views |
check_repository |
AST rule engine + security audit |
explain_model |
Rich model explanation with knowledge base |
list_models |
List all models in repository |
list_views |
List all views in repository |
list_actions |
List all actions in repository |
repository_summary |
High-level repository stats |
models_missing_acl |
Security audit: models without ACL |
inheritance_graph |
Model inheritance graph as Mermaid |
build_index |
Rebuild repository index |
list_knowledge_topics |
List domain knowledge topics |
Example: AI-assisted debugging
User: "Why does sale.order have no ACL?"
→ Agent calls `search_model("sale.order")` + `models_missing_acl`
→ "Model sale.order is defined at models/sale.py:10 but has no ACL entry in security/ir.model.access.csv"
CLI Reference
# Full analysis with markdown report
python -m analyzer.cli path/to/addon
# JSON output for CI pipelines
python -m analyzer.cli path/to/addon --format json
# SARIF for GitHub Code Scanning
python -m analyzer.cli path/to/addon --format sarif -o results.sarif
# Quick check (violations only)
python -m analyzer.cli path/to/addon --check
# Check with baseline suppression (legacy repos)
python -m analyzer.cli path/to/addon --check --baseline
# Generate baseline from existing violations
python -m analyzer.cli path/to/addon --baseline
# Build searchable index for AI tools
python -m analyzer.cli path/to/addon --index -o repo_index
# Rule statistics
python -m analyzer.cli path/to/addon --stats
# Inheritance graph (Mermaid format)
python -m analyzer.cli path/to/addon --graph
Rule Statistics
python -m analyzer.cli path/to/addon --stats
Output:
Repository: my_addon
Models: 12
Fields: 142
Views: 8
Actions: 5
ACLs: 10
Record Rules: 3
Violations by severity:
CRITICAL: 2
HIGH: 7
MEDIUM: 13
LOW: 21
─────────────────
Total: 43
Top rules violated:
orm-no-n-plus-1 (CRITICAL) → 3 violations
orm-raw-sql (HIGH) → 2 violations
security-acl (CRITICAL) → 2 violations
Confidence Score
Each violation includes a confidence score to help prioritize review:
search-inside-loop
Severity: CRITICAL
Confidence: 98% ← High accuracy → fix immediately
missing-index
Severity: MEDIUM
Confidence: 60% ← Possible false positive → review needed
| Confidence | Meaning | Action |
|---|---|---|
| 90-100% | Static analysis certainty | Fix immediately |
| 70-89% | Clear pattern match | Quick review |
| 50-69% | Heuristic-based | Needs verification |
| < 50% | Low confidence | May ignore |
Use Cases
1. Code Review Automation
GitHub Action automatically reviews every PR:
# .github/workflows/odoo-review.yml
- name: Odoo Review
uses: ./.github/actions/odoo-review
with:
fail-on-critical: 'true'
2. Legacy Repository Onboarding
# Step 1: Baseline existing violations
python -m analyzer.cli /path/to/legacy/addon --baseline
# → Creates odoo-baseline.json with 387 accepted violations
# Step 2: Only NEW violations are reported from now on
python -m analyzer.cli /path/to/legacy/addon --check --baseline
# → Only 3 new violations remain
3. AI-assisted Development
MCP server enables AI agents to understand your Odoo codebase:
→ "Explain model sale.order"
→ Agent returns: fields, methods, views, inheritance chain, knowledge docs
4. CI Pipeline Quality Gate
# Fail build if CRITICAL or HIGH violations exist
- name: Quality Gate
run: |
python -m analyzer.cli . --check --format json -o report.json
python -m ci_tools/check_gate.py report.json
Baseline System
Enables onboarding legacy repositories without being overwhelmed by existing violations.
# Generate baseline
python -m analyzer.cli /path/to/addon --baseline
# Check with baseline
python -m analyzer.cli /path/to/addon --check --baseline
Baseline data is stored as odoo-baseline.json:
{
"version": 1,
"addon": "my_addon",
"timestamp": "2026-06-20T12:00:00Z",
"total_accepted": 387,
"accepted": [
{"rule": "orm-no-n-plus-1", "file": "models/foo.py", "line": 54}
]
}
Programmatic API
from analyzer import RepositoryStore
store = RepositoryStore("path/to/addon")
store.load()
# Model queries
store.search_model("sale.order")
store.fields_for_model("sale.order")
store.methods_for_model("sale.order")
# Analysis
store.check_code()
store.violations_by_severity("CRITICAL")
# Security audit
store.models_missing_acl()
store.list_acls()
# Graphs
print(store.inheritance_graph_mermaid())
# Summary
print(store.repository_summary())
Project Structure
odoo-best-practices/
├── SKILL.md # Entry point — rule index + triggers
├── AGENTS.md # Full compiled reference (136 rules)
├── README.md # This file
├── rules/ # 136 rule files (13 categories)
├── bad_patterns/ # 12 anti-pattern detectors
├── knowledge/ # 12 core model domain files
├── versions/{14,15,16,17,18,19}/ # Version-specific guides
├── docs/ # 60 official Odoo docs (14-19)
├── analyzer/ # Python static analysis engine
│ ├── cli.py # CLI: 7 sub-commands
│ ├── checker.py # AST rule engine (4 checks)
│ ├── constants.py # Shared severity constants
│ ├── indexer.py # Repository index builder
│ ├── reporter.py # Markdown/JSON/SARIF report generator
│ ├── baseline.py # Baseline suppression system
│ ├── sarif.py # SARIF v2.1.0 output with fix suggestions
│ ├── store.py # RepositoryStore (unified API)
│ ├── mcp_server.py # MCP protocol server (14 tools)
│ ├── graph.py # Inheritance + dependency graphs
│ ├── init_generator.py # Per-IDE config generator (5 IDEs)
│ └── parsers/
│ ├── common.py # Shared AST helpers (ast_node_to_value)
│ ├── manifest_parser.py # AST-safe manifest parsing
│ ├── model_parser.py # Models, fields, methods, decorators
│ ├── view_parser.py # Views, actions, menus, templates
│ └── security_parser.py # ACLs, record rules, groups, categories
├── package.json # npm package (scoped @foxpink-dev)
├── bin/ # npm CLI entry point
├── .npmignore
├── .github/
│ ├── workflows/odoo-review.yml # PR review CI
│ ├── workflows/publish-npm.yml # npm publish
│ └── actions/odoo-review/ # Docker action for GitHub CI
│ └── entrypoint.py
Version Support
| Version | Status | Key Features |
|---|---|---|
| 14 | ✅ Legacy | Pre-OWL, Classic Widgets |
| 15 | ✅ Legacy | OWL 1.x introduced |
| 16 | ✅ Stable | OWL 2.x Default |
| 17 | ✅ Stable | OWL Required, t-out |
| 18 | ✅ Current | Legacy JS removed, OWL 3 |
| 19 | ✅ Latest | <list> tag canonical |
License
MIT
Portfolio Note
Built an Odoo Engineering Platform featuring repository intelligence, static analysis, MCP integration, GitHub-native code review, SARIF reporting, baseline suppression, and AI-assisted development workflows for Odoo 14–19.
This project demonstrates:
- AST parsing — Python source code analysis
- Static analysis — Rule engine, pattern detection
- DevSecOps — CI/CD integration, quality gates
- MCP/AI tooling — 14 tools for AI agents
- Rule engine design — 136 rules across 13 categories
- Platform engineering — From CLI to GitHub Action to AI integration
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。