ai-automation-workflows

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill implements patterns that ingest raw file content (cat $file) and command-line arguments into LLM prompts. Without explicit boundary markers or sanitization, these scripts are susceptible to indirect prompt injection if the processed data contains adversarial instructions.\n
  • Ingestion points: data_processing.sh ($(cat $file)), conditional_workflow.sh ($INPUT_TEXT).\n
  • Boundary markers: Absent.\n
  • Capability inventory: File system access, network operations (via curl), and CLI tool execution.\n
  • Sanitization: Absent.\n- [EXTERNAL_DOWNLOADS]: The documentation references external installation scripts and additional skills hosted on inference.sh infrastructure and associated GitHub repositories for the platform's toolset.\n- [COMMAND_EXECUTION]: The skill uses Bash to orchestrate multi-step workflows, including file system management (mkdir), process control (wait, &), and system utility modification (crontab) for scheduling tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 06:04 PM
Security Audit — agent-trust-hub — ai-automation-workflows