rw-generate-image
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill accepts external reference images via URLs, creating a vulnerability surface where adversarial image data could attempt to influence the agent's logic or downstream tasks.
- Ingestion points: Remote URLs provided via the
--reference-imagesparameter inSKILL.md. - Boundary markers: No specific delimiters or boundary protections are implemented for the injected URLs.
- Capability inventory: The skill has the ability to perform network requests to the Runway API and write image files to the local file system.
- Sanitization: The documentation notes reliance on the API's internal safety filters (e.g.,
SAFETY.INPUT.*). - [COMMAND_EXECUTION]: The skill executes local Python scripts using the
uv runcommand. While restricted to thescripts/directory, this mechanism executes logic that interacts with both the local filesystem and the network.
Audit Metadata