thirdparty-ui-ux-pro-max
Warn
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: MEDIUMINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides an interface that processes untrusted input alongside file-writing capabilities.
- Ingestion points: User input is ingested through the
queryargument inscripts/search.pyand theproject_nameandpageparameters inscripts/design_system.py. - Boundary markers: Absent. Search results from the CSV database are returned to the agent without delimiters or instructions to disregard potentially malicious instructions embedded in the data.
- Capability inventory: The skill allows writing files to the local filesystem via the
persist_design_systemfunction. - Sanitization: Minimal sanitization is performed on user inputs used to determine file storage locations.
- [COMMAND_EXECUTION]: The skill executes local Python scripts that handle file operations unsafely.
- Evidence: In
scripts/design_system.py, the functionpersist_design_systemconstructs file paths using theproject_nameandpagevariables. These variables are transformed into slugs using.lower().replace(' ', '-'), which fails to remove directory traversal sequences like../. This allows an attacker or a malicious prompt to manipulate the path and writeMASTER.mdor other.mdfiles to arbitrary locations on the disk, potentially overwriting existing documentation or project files.
Audit Metadata