article-to-cover
Warn
Audited by Gen Agent Trust Hub on Mar 19, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script (
run_command.py) via a shell command in Step 4 ofSKILL.md. It dynamically constructs a command string that includes apromptgenerated from user-influenced data. This pattern is susceptible to argument injection if the input data contains characters that escape the JSON or shell quoting context. - [PROMPT_INJECTION]: The skill exhibits a significant surface for Indirect Prompt Injection.
- Ingestion points: The skill reads and analyzes untrusted data from articles, chat transcripts, and design briefs provided by the user (Step 1,
SKILL.md). - Boundary markers: There are no explicit delimiters or instructions to the agent to ignore embedded commands within the processed text.
- Capability inventory: The skill has the ability to execute shell commands to perform image generation tasks (Step 4,
SKILL.md). - Sanitization: The skill lacks mechanisms to sanitize or validate the external text before it is used to determine design directions and generate production instructions.
- [EXTERNAL_DOWNLOADS]: The skill relies on the
meitu-aiengine and associated scripts for its core functionality. While Meitu is a well-known service, the skill's reliance on a script located at a relative path (../meitu-ai/scripts/run_command.py) outside its own directory represents dynamic loading from a computed path.
Audit Metadata