privacy-webhook-analyzer
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a potential surface for indirect prompt injection due to its data processing nature.
- Ingestion points: The
run.pyscript reads input from a user-specified file path usingPath.read_text(). - Boundary markers: No specific boundary markers or instructions are provided to the agent to treat the input data as untrusted content.
- Capability inventory: The script allows for reading from and writing to the local filesystem via command-line arguments.
- Sanitization: The implementation uses standard JSON parsing but does not include explicit logic to sanitize or filter potential instructions embedded within the input data.
Audit Metadata