wechat-sticker-maker
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions in SKILL.md require the agent to analyze the visual content of processed images to automatically generate and fill metadata files (info.txt and meta.txt). This creates a risk where malicious instructions embedded in images could be interpreted by the agent, potentially leading to the generation of unintended metadata content or influencing subsequent agent actions. Ingestion points: The user-provided source grid image file. Boundary markers: Absent; the skill does not include instructions to the agent to treat image content as untrusted or ignore embedded textual instructions. Capability inventory: The skill performs file creation and writing operations in the output directory and image processing. Sanitization: Absent; there is no validation or filtering of the agent-interpreted text before it is written to the metadata files.
- [COMMAND_EXECUTION]: The scripts/run.sh script automates the creation of a local Python virtual environment and installs the Pillow library, which is a standard and safe dependency for image processing. These operations are confined to the skill directory and use official package registries.
Audit Metadata