CyberSentinel AI is an automated security monitoring and AI analysis system. It watches CVE-related repositories, security-focused GitHub repositories, and security articles, then produces local Markdown/HTML reports. Blog publishing is optional.

Features

Monitor GitHub for CVE/POC-related repositories.
Search GitHub by security keywords.
Track configured high-priority repositories.
Analyze findings with OpenAI-compatible providers and Gemini.
Store monitor state in SQLite.
Generate Markdown and HTML reports.
Publish reports to a blog API only when BLOG_TOKEN is configured.
Generate GitHub Actions Markdown report artifacts without committing runtime data.

Quick start

git clone https://github.com/Hxnxe/CyberSentinel-AI.git
cd CyberSentinel-AI

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

cp .env.example .env
python main.py --check-config
python -m pytest -q
python main.py

The project can be imported and tested without private credentials. Add your own .env values before real monitoring.

You can also use the bootstrap script:

bash scripts/bootstrap.sh

Minimal production configuration

GITHUB_TOKENS=["replace-with-your-github-token"]
PRIMARY_AI_KEY=replace-with-your-openai-compatible-key
PRIMARY_AI_BASE=https://api.openai.com/v1
PRIMARY_AI_MODEL=gpt-4.1-nano

Optional providers and integrations:

BACKUP_AI_CONFIGS=[]
GEMINI_AI_KEY=
BLOG_BASE_URL=http://127.0.0.1
BLOG_TOKEN=

When BLOG_TOKEN is empty, the application skips blog API calls and keeps reports local.

Commands

Run tests:

python main.py --check-config
python -m pytest -q

Start the monitor:

python main.py

Install as an editable package:

pip install -e .
cybersentinel-ai

Project layout

ai/          AI clients, prompts, schemas, validation, and analysis
clients/     GitHub API client
database/    SQLAlchemy models, sessions, repositories, initialization
monitors/    CVE and GitHub repository monitors
reports/     Markdown/HTML report generation
utils/       logging, article, blacklist, blog, CSV, and Markdown helpers
scripts/     bootstrap and configuration doctor helpers
docs/        configuration, quick-start, and deployment notes
tests/       pytest suite
main.py      runtime entrypoint
config.py    environment-backed configuration

Private and generated files

The release excludes local secrets and runtime artifacts:

.env
.env.*
venv/
.venv/
logs/
results/
data/
cve_database.db
*.db
*.sqlite
__pycache__/
.pytest_cache/

Never commit real tokens, generated databases, logs, reports, runtime data, or virtual environments.

Configuration reference

Variable	Purpose
`MONITOR_INTERVAL`	Monitoring interval in seconds. Default: 7200
`DATABASE_PATH`	SQLite database path. Default: `cve_database.db`
`GITHUB_TOKENS`	JSON array of GitHub tokens
`GITHUB_TOKEN`	Single fallback GitHub token
`PRIMARY_AI_KEY`	Primary AI provider API key
`PRIMARY_AI_BASE`	Primary AI provider base URL
`PRIMARY_AI_MODEL`	Primary AI model
`BACKUP_AI_CONFIGS`	JSON array of backup OpenAI-compatible providers
`GEMINI_AI_KEY`	Gemini API key
`BLOG_BASE_URL`	Blog API base URL
`BLOG_TOKEN`	Blog API token. Leave empty to disable blog publishing

License

MIT. See LICENSE.