workspace-seat-utilization-auditor
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a standard Python scaffold for data processing. It reads from a user-specified input file and writes to a user-specified output file using standard libraries. No unauthorized network access, remote code execution, or credential exposure patterns were found.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external data (JSON/CSV), which is a common vector for indirect prompt injection. However, this is consistent with its primary purpose as an auditing tool.\n
- Ingestion points: The
run.pyscript reads from a file path provided through the--inputcommand-line argument.\n - Boundary markers: None present (standard for a CLI scaffold).\n
- Capability inventory: The script has capabilities to read and write files on the local file system.\n
- Sanitization: The skill uses standard
json.loadsfor parsing input data, which does not sanitize instructions embedded in values.
Audit Metadata