deepscan-create-task
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to execute a local Python script (
scripts/create_task.py) to process user requests. - [DATA_EXFILTRATION]: The script reads a stored authentication token from
~/.deepscan/tokenand transmits it via an Authorization header to the vendor's API atdata.cli.im. This is a necessary step for authenticated task creation on the platform. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by interpolating user-provided inputs (such as the scanning book's name and description) directly into command-line arguments in
SKILL.md. Although the Python script employsargparseto handle these inputs, the lack of explicit boundary markers or input sanitization in the markdown instructions presents a potential surface for argument manipulation.
Audit Metadata