service-ownership-mapper
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The provided code is a standard implementation for local data processing. It relies on Python's standard library to read input from a file and write a summary to an output file. No external packages, remote downloads, or obfuscated code are present.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data, which is a potential surface for indirect prompt injection. However, the implementation is safe as it only sorts and returns keys, discarding values that might contain malicious instructions. * Ingestion points: The run.py script reads a user-provided file via the --input flag. * Boundary markers: None identified in SKILL.md. * Capability inventory: Local file system read/write access via pathlib.Path in run.py. * Sanitization: The script performs JSON parsing but does not specifically sanitize input content.
Audit Metadata