ai-automation-workflows
Warn
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DATA_EXFILTRATION]: The monitoring script
monitored_workflow.shdemonstrates sending execution results and error logs to an external URL usingcurl. This provides a mechanism for data to be transmitted outside the local environment. - [COMMAND_EXECUTION]: The skill documentation includes instructions for setting up persistent automation using
crontab, allowing scripts to run on a recurring schedule. It also utilizessubprocess.runin Python and extensive Bash scripting to interact with the system and theinfshCLI. - [PROMPT_INJECTION]: The
data_processing.shtemplate demonstrates an indirect prompt injection surface. Ingestion points: The script reads raw content from local text files in./data/raw/. Boundary markers: Absent. Capability inventory: Use ofinfsh app runfor AI tasks andcurlfor network operations across scripts. Sanitization: Absent; content is used without validation or escaping. - [EXTERNAL_DOWNLOADS]: The documentation suggests installing dependencies using
npx, which downloads and executes packages from external registries.
Audit Metadata