ai-automation-workflows

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides numerous Bash and Python scripts that execute commands via the infsh CLI to manage AI workflows.
  • [PROMPT_INJECTION]: The data_processing.sh template presents an indirect prompt injection surface. [Ingestion points]: Reads data from ./data/raw/*.txt files. [Boundary markers]: Missing; file content is directly interpolated into the LLM prompt. [Capability inventory]: Executes LLM tasks and writes files to the local system. [Sanitization]: None; the content is passed directly from cat to the prompt.
  • [DATA_EXFILTRATION]: Includes a monitoring example that demonstrates how to send error alerts to an external webhook, using a placeholder URL (https://your-webhook.com/alert).
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 02:29 PM