scenario-product-shots
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from two main sources: user-provided text descriptions and uploaded image assets.
- Ingestion points: User-supplied words are passed to the
recommendtool; images uploaded viaupload_assetare processed byasset_analyzeand used as references for generative models. - Boundary markers: The instructions recommend using quotes for label text within prompts ("the label reads 'SUMMIT COLD BREW'"), but there are no explicit system instructions provided to the agent to ignore potential malicious commands embedded within the text or images.
- Capability inventory: The skill uses a variety of tools from the Scenario ecosystem (
upload_asset,asset_download,asset_analyze,recommend) which involve network operations and image processing/generation. - Sanitization: No specific sanitization, filtering, or validation steps for external content are defined in the instructions.
- [EXTERNAL_DOWNLOADS]: The skill mentions installing sibling skills from the author's repository using
npx skills add scenario-labs/skills. This involves fetching external code from the vendor's GitHub organization. - [COMMAND_EXECUTION]: The skill references a shell command (
npx skills add) to be recommended to the user if specific sibling skills are missing. This keeps the execution within the user's control (human-in-the-loop).
Audit Metadata