naming-convention
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external metadata from JSON files to generate filenames, creating a surface for indirect prompt injection if the source data is attacker-controlled.\n
- Ingestion points: The
scripts/rename_batch.pyscript reads data from files provided via the--mapand--manifestarguments.\n - Boundary markers: Data is parsed as structured JSON; however, no explicit instructions are provided to the model to ignore instructions embedded in the metadata.\n
- Capability inventory: The script performs file system modifications including renaming files (
os.rename) and updating manifest files.\n - Sanitization: Filename components are processed through a
slug()function that restricts values to lowercase alphanumeric characters and hyphens, mitigating path traversal or shell character injection via the data values.\n- [COMMAND_EXECUTION]: The skill includes a Python script that executes file system operations to rename deliverables. The script is designed for safety, incorporating a dry-run default and checks to prevent file clobbering or name collisions.
Audit Metadata