gemini-carousel
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses directive language ('CRITICAL: Auto-start on load', 'When this skill triggers, go straight to Step 1. Do not summarise.') intended to force the agent to bypass its standard conversational behavior or summary routines.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to read and process content from external files within the project environment (specifically 'brand-kit.md' and 'voice.md').
- Ingestion points: Reads content from
brand-kit.mdandvoice.mdto determine styling and language preferences. - Boundary markers: None identified; the skill directly interpolates content from these files into its logic.
- Capability inventory: The skill primarily outputs text and structured image generation prompts to the user; it does not appear to have file-write or direct network capabilities.
- Sanitization: There is no evidence of sanitization or filtering for the content read from the external project files, which could lead to instructions in those files influencing the agent's behavior.
Audit Metadata