taxonomy-consistency-linter
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from local files (JSON/CSV) and returns it to the agent, which presents a surface for indirect prompt injection if the processed data contains malicious instructions. * Ingestion points: The main entry point in run.py (line 29) reads input from a provided file path. * Boundary markers: None are specified to protect the agent from embedded instructions. * Capability inventory: The skill is limited to local file read and write operations based on CLI arguments. * Sanitization: The skill performs standard JSON parsing but no specific content sanitization.
- [SAFE]: The implementation is a boilerplate reference scaffold using only standard Python libraries (argparse, json, pathlib) with no external network activity or obfuscated code.
Audit Metadata