boogu-image
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEDATA_EXFILTRATIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DATA_EXFILTRATION]: The skill instructs the agent to access
~/.pixazo/api-keyand thePIXAZO_API_KEYenvironment variable. These credentials are transmitted togateway.pixazo.aito authenticate image generation requests. While this is standard for the service, it involves reading sensitive file paths and network transmission to a non-whitelisted domain.\n- [REMOTE_CODE_EXECUTION]: The skill provides instructions to executenpx skills add Pixazo-AI/skills, which downloads and runs code from a remote source. It also suggests fetching instructions fromhttps://www.pixazo.ai/models/boogu-image.md.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided prompts and reference images which are then interpolated into API requests.\n - Ingestion points: User prompts and reference image data processed in SKILL.md.\n
- Boundary markers: Absent; there are no instructions to delimit user input or ignore embedded commands.\n
- Capability inventory: Network request capabilities via
curl,requests, andfetchtools.\n - Sanitization: Absent; the skill does not define validation or filtering for user-provided inputs.\n- [EXTERNAL_DOWNLOADS]: The skill references external resources for configuration and related skill installation, targeting
www.pixazo.aiand thePixazo-AI/skillsrepository.
Audit Metadata