ds-pro-max
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION] (MEDIUM): The
scripts/generate.pytool accepts a user-defined--outputfile path and writes content to it usingopen(output_path, 'w'). The script lacks path validation or sanitization, creating a risk of path traversal where a malicious prompt could trick the agent into overwriting critical system or configuration files (e.g.,.bashrcor.ssh/config). - [COMMAND_EXECUTION] (LOW): The
SKILL.mdfile provides setup instructions that includesudo apt install. While intended for legitimate environment setup, recommending the use of administrative privileges is a high-risk pattern that can be misused if the surrounding environment is not properly isolated. - [INDIRECT PROMPT INJECTION] (LOW): The skill is susceptible to indirect injection through its processing of design specifications and user queries.
- Ingestion points: User-provided query strings in
search.pyandgenerate.py. - Boundary markers: Absent. Instructions do not specify delimiters for user-provided data when passed to scripts.
- Capability inventory: Local file writing in
generate.py. - Sanitization: None. Data from queries is interpolated into templates or used for file naming without validation.
Audit Metadata