aiguard-safety

v0.7.5.9 suspicious
8.0
High Risk

AIGuard: model-agnostic safety evaluation toolkit with PDF reports (adversarial, evaluator, hallucination)

🤖 AI Analysis

Final verdict: SUSPICIOUS

The package aiguard-safety v0.7.5.9 has significant metadata risks due to low activity and missing maintainer information, raising suspicion of potential malicious intent.

  • Low activity and lack of maintainer details in metadata.
  • Potential signs of a supply-chain attack.
Per-check LLM notes
  • Metadata: The low activity and lack of maintainer details raise concerns about potential malicious intent.

📦 Package Quality Overall: Medium (5.2/10)

◈ Medium Test Suite 6.0

Partial test coverage signals detected

  • 2 test file(s) detected (e.g. base_test.py)
◈ Medium Documentation 5.0

Some documentation present

  • Detailed PyPI description (40609 chars)
○ Low Contributing Guide 4.0

No contributing guide or governance files found

  • Development Status classifier >= Beta
◈ Medium Type Annotations 5.0

Partial type annotation coverage

  • 215 type-annotated function signatures detected in source
◈ Medium Multiple Contributors 6.0

Limited contributor diversity

  • 2 unique contributor(s) across 92 commits in Shelton03/aiguard
  • Two distinct contributors found

🔬 Heuristic Checks

⚠ Outbound Network Calls score 4.5

Found 3 network call pattern(s)

  • rer {api_key}" req = urllib.request.Request(endpoint, data=data, headers=headers, method="POST")
  • try: with urllib.request.urlopen(req, timeout=self._config.timeout_s) as resp:
  • arer {api_key}" req = urllib.request.Request(endpoint, data=data, headers=headers, method="POST")
✓ Code Obfuscation

No obfuscation patterns detected

⚠ Shell / Subprocess Execution score 6.0

Found 3 shell execution pattern(s)

  • s…") try: subprocess.run([npm_cmd, "install"], cwd=str(ui_dir), check=False)
  • production bundle…") subprocess.run([npm_cmd, "run", "build"], cwd=str(ui_dir), check=False)
  • .cmd" try: proc = subprocess.Popen( [npm_cmd, "run", "preview", "--", "--port", str
✓ Credential Harvesting

No credential harvesting patterns detected

✓ Typosquatting

No typosquatting candidates detected

✓ Registered Email Domain

Email domain looks legitimate: gmail.com>

✓ Suspicious Page Links

All external links appear legitimate

⚠ Git Repository History score 2.5

Git history flags: Repository has zero stars and zero forks

  • Repository has zero stars and zero forks
⚠ Maintainer History score 4.0

2 maintainer concern(s) found

  • Author name is missing or very short
  • Author "" appears to have only 1 package on PyPI (new or inactive account)
✓ Known CVE Vulnerabilities

No known vulnerabilities found in OSV database.

💡 AI App Starter Prompt

Use this prompt to build a project with aiguard-safety 
Develop a comprehensive AI safety evaluation tool called 'SafetyGuard' using the Python package 'aiguard-safety'. This tool will serve as a user-friendly interface for evaluating the safety of any AI model based on adversarial attacks, evaluator performance, and hallucination detection. The application should generate detailed PDF reports summarizing the findings.

Step 1: Setup the Project Environment
- Install necessary libraries including 'aiguard-safety', 'pandas', 'matplotlib', and 'reportlab' for PDF generation.
- Create a virtual environment and activate it.

Step 2: Design the User Interface
- Use a simple and intuitive graphical user interface (GUI) framework such as Tkinter or PyQt5.
- Implement input fields for users to upload their AI models.
- Include options for selecting the type of safety checks (adversarial, evaluator, hallucination).

Step 3: Integrate 'aiguard-safety'
- Utilize 'aiguard-safety' to perform adversarial attacks on the uploaded AI models.
- Evaluate the robustness of the models against these attacks.
- Detect potential hallucinations produced by the models under different scenarios.
- Assess the evaluator's performance in accurately identifying safe versus unsafe outputs.

Step 4: Generate Detailed Reports
- Compile the results from each safety check into a structured format.
- Use 'reportlab' to create professional-looking PDF reports that include charts, graphs, and textual summaries.
- Ensure the reports are easy to understand and provide actionable insights.

Suggested Features:
- Real-time progress indicators during the safety evaluations.
- Option to save and load previous model evaluations.
- Comparative analysis feature allowing side-by-side comparisons of different models.
- Customizable threshold settings for defining safe vs. unsafe behavior.
- Integration with cloud storage services for backing up and sharing reports.