Human-in-the-loop agentic AI CLI for penetration testers that runs recon, validation, and evidence-backed reporting in the terminal.
zakirkun/guardian-cli
A Python CLI that uses LLMs such as Google Gemini, OpenAI, and Claude to orchestrate penetration-testing workflows around established security tools like nmap and nuclei.
- Stars
- 1,895
- Forks
- 368
- Open issues
- 5
- Last commit
- 4 months ago
What is it?
Guardian is a command-line framework for automating penetration tests. An AI agent plans the testing steps, selects which security tools to run, and analyzes their output, while every tool execution is recorded so each finding can be traced back to the command that produced it.
It supports several AI providers, including OpenAI, Claude, Gemini, OpenRouter, Requesty, Ollama, and OpenAI-compatible endpoints. The project ships about 50 tool wrappers covering network scanning, web reconnaissance, subdomain discovery, vulnerability scanning, SSL/TLS testing, cloud and container audits, and Active Directory and Android assessments. Workflows are defined in YAML as a dependency graph, and steps can run in parallel and resume after interruption.
Reports can be generated as Markdown, HTML, or JSON, and findings can be exported as SARIF, to DefectDojo, or to Slack. The tool includes scope validation, a confirmation gate for intrusive tools, and redaction of API keys in logs. Guardian is intended only for systems the user is authorized to test.
Key features
- 01Multi-provider AI planning and tool selection, including Gemini, OpenAI, Claude, and local Ollama models.
- 02About 50 wrapped security tools across network, web, cloud, and mobile categories.
- 03YAML workflows with dependency-based parallel execution and resume support.
- 04Evidence capture that links each finding to its source tool execution.
- 05Report export in Markdown, HTML, JSON, and SARIF, plus DefectDojo and Slack integrations.
- 06Scope validation and a confirmation gate for intrusive tools.
Use cases
- Running a repeatable reconnaissance and web vulnerability scan against systems you are authorized to test.
- Producing structured pentest reports with evidence and export to issue trackers or SARIF-based code scanning.
- Experimenting with LLM-driven security automation using a local model via Ollama.
Quick start
- 1
Clone the repository
git clone https://github.com/zakirkun/guardian-cli.git - 2
Enter the project directory
cd guardian-cli - 3
Create a virtual environment and install (Linux/macOS)
python3 -m venv venv && source venv/bin/activate && pip install -e . - 4
Verify the installation
python -m cli.main --help
Things to know
- Requires an API key from a supported AI provider such as OpenAI, Anthropic, Google, OpenRouter, or Requesty.
- Many security tools are optional and must be installed separately for full functionality.
- Licensed under NOASSERTION in metadata, although the README states MIT; check the LICENSE file before use.
- Intended only for authorized security testing, so use carries legal responsibility.
Topics
FAQ
- What is guardian-cli?
- A Python CLI that uses LLMs such as Google Gemini, OpenAI, and Claude to orchestrate penetration-testing workflows around established security tools like nmap and nuclei. Guardian is a command-line framework for automating penetration tests. An AI agent plans the testing steps, selects which security tools to run, and analyzes their output, while every tool execution is recorded so each finding can be traced back to the command that produced it.
- Is guardian-cli free to use? What license does it have?
- guardian-cli is open source under the Other license. Check the license terms before using it in commercial products.
- Is guardian-cli still maintained?
- Updated within 6 months. The last commit to guardian-cli was on Jun 27, 2026.
- What can I use guardian-cli for?
- Running a repeatable reconnaissance and web vulnerability scan against systems you are authorized to test. Producing structured pentest reports with evidence and export to issue trackers or SARIF-based code scanning. Experimenting with LLM-driven security automation using a local model via Ollama.
- How do I install guardian-cli?
- Clone the repository: git clone https://github.com/zakirkun/guardian-cli.git Enter the project directory: cd guardian-cli Create a virtual environment and install (Linux/macOS): python3 -m venv venv && source venv/bin/activate && pip install -e . Verify the installation: python -m cli.main --help
- Where can I get guardian-cli?
- The source code and installation instructions are on GitHub: https://github.com/zakirkun/guardian-cli
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A Python CLI that automatically locks your desktop when your paired Bluetooth phone or smartwatch moves out of range.
A self-improving AI agent that creates and refines skills from experience, runs on many model providers, and works through CLI and messaging apps.
A toolkit that generates agent-native command-line interfaces for existing software, plus a CLI-Hub registry for installing ready-made ones.
A command-line tool that gives one interface to Google Workspace APIs like Drive, Gmail, Calendar, Sheets, Docs, and Chat, with JSON output for scripts and AI agents.
A fast, cross-platform task runner written in Go, inspired by Make and configured through a Taskfile YAML file.