workflow-ecommerce-launch
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes multiple local Python scripts (dfs_client.py, schema_gen.py, and report_publish.py) via shell commands throughout its workflow steps.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a vulnerability surface where user-provided input is used in command-line arguments.
- Ingestion points: User-supplied 'category name' and 'product type' strings from SKILL.md.
- Boundary markers: No specific boundary markers or instructions to treat inputs as raw data are provided for the shell command interpolation, although inputs are enclosed in double quotes in the examples.
- Capability inventory: The skill utilizes subprocess execution of Python scripts (scripts/dfs_client.py, scripts/schema_gen.py, scripts/report_publish.py) which could be exploited for arbitrary command execution.
- Sanitization: No explicit sanitization, filtering, or validation of user-provided inputs is mentioned in the skill instructions to prevent command termination or injection.
Audit Metadata