python-backend
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_CONTEXT_INJECTION
Full Analysis
- [DYNAMIC_CONTEXT_INJECTION]: The 'scripts/create-fastapi-app.md' file utilizes shell command substitution (the '!command' syntax) to perform local environment discovery, such as identifying the Python version, Git repository root, and current project dependencies. These commands execute at load time to populate template variables. Analysis shows these commands are focused on benign project reconnaissance and do not involve network operations, sensitive file access (e.g., SSH/AWS keys), or direct user input injection.
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting external configuration data. 1. Ingestion points: 'scripts/create-fastapi-app.md' reads local files 'pyproject.toml' and 'requirements.txt' to customize templates. 2. Boundary markers: No explicit instructions to ignore embedded commands are present in the detection logic. 3. Capability inventory: The skill has 'Read' and 'Grep' tool access and provides code generation templates. 4. Sanitization: The skill implements mitigation through restrictive patterns, using 'grep -oE' and 'awk' to extract only known library names and 'head' to limit the amount of processed text, which reduces the potential for malicious data to influence the agent's behavior.
Audit Metadata