robot-framework-py

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references several well-known and official Robot Framework libraries for installation via pip, including 'robotframework-seleniumlibrary', 'robotframework-requests', 'robotframework-browser', 'robotframework-appiumlibrary', and 'robotframework-databaselibrary'. These are standard industry tools for test automation.
  • [DATA_EXPOSURE]: The skill explicitly instructs against hardcoding secrets. Examples like 'examples/requests/variables/env.py' correctly demonstrate fetching sensitive data like 'API_TOKEN' from environment variables using 'os.getenv', adhering to secure development practices.
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as it is designed to generate executable Robot Framework code from user input. However, the author has mitigated this risk by including a comprehensive 'acceptance-criteria.md' file and 'scenarios.yaml' that enforce 'Do-not-hallucinate' constraints, forbidden patterns (like 'eval' or hardcoded passwords), and strict syntax validation.
  • Ingestion points: User prompts requesting the generation of .robot suites, .resource files, or custom Python keyword libraries.
  • Boundary markers: The skill uses 'expected_patterns' and 'forbidden_patterns' in its training scenarios to define output boundaries.
  • Capability inventory: The skill generates automation scripts and Python files capable of executing web, API, and system operations.
  • Sanitization: Strict enforcement of library imports and forbidden keyword patterns reduces the risk of malicious code generation.
  • [SAFE]: All external documentation links, such as those pointing to 'arvind3.github.io', 'deepwiki.com', and 'gitdiagram.com', are used appropriately for project documentation and architecture visualization.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 05:52 AM