ideogram-deploy-integration
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements an API endpoint that processes untrusted user data (image prompts, styles, and aspect ratios) which is then forwarded to the Ideogram API.
- Ingestion points: The
app/api/generate/route.tsfile reads data from the JSON request body. - Boundary markers: No explicit delimiters or instructions are used to prevent the Ideogram API from interpreting malicious instructions inside the prompt string.
- Capability inventory: The endpoint performs network operations (fetching from
api.ideogram.aiand uploading to S3 storage). - Sanitization: The code includes a basic length check for the prompt (
prompt.length > 10000) and validates the prompt's existence.
Audit Metadata