krea-generate
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill performs a version check by fetching data from the vendor's official GitHub repository (
raw.githubusercontent.com/krea-ai/skills/main/VERSION) viacurl. This is a routine update check from a trusted source. - [COMMAND_EXECUTION]: The skill instructions direct the agent to execute a local bash script (
scripts/update-check.sh) to perform the aforementioned version check and manage local state in~/.krea-skills. - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and processes untrusted external media.
- Ingestion points: User-provided images and external media URLs defined in
SKILL.md(Universal Rule 7). - Boundary markers: None explicitly implemented for external content processing.
- Capability inventory: The skill utilizes various image and video generation tools through Krea MCP, including
openai/gpt-image-2,google/nano-banana-pro, andbytedance/seedance-2-5. - Sanitization: No specific sanitization or filtering of external image metadata or content is described.
Audit Metadata