codex-pet
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local commands including
runcomfyandmagick(ImageMagick) to perform image transformations and assemble spritesheet atlases. These operations are consistent with the skill's stated purpose. - [EXTERNAL_DOWNLOADS]: The pipeline fetches source images from external, user-provided URLs to serve as the visual basis for the generated mascot.
- [PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface where the source image provided by the user is processed by an image generation model (GPT Image 2). Malicious instructions could theoretically be embedded within the image data.
- Ingestion points: The
SOURCE_URLvariable in the generation script withinSKILL.mdcaptures external data. - Boundary markers: Input is passed to the CLI as a string; no explicit sanitization is performed on the image content itself.
- Capability inventory: The skill has the ability to write to the local filesystem (mkdir, cp, cat) and execute image processing commands.
- Sanitization: No local sanitization is performed on the image; the skill relies on the safety guardrails of the remote RunComfy/OpenAI API.
Audit Metadata