kuroco-api-content
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent on how to retrieve and process content from external API endpoints, which creates a potential surface for indirect prompt injection if the content within the CMS is malicious.
- Ingestion points: External data is fetched using the
fetch()API and parsed as JSON inSKILL.md,references/file-operations.md, andreferences/filter-query.md. - Boundary markers: The instructional code examples do not include delimiters or specific prompts to ignore instructions that might be embedded in the fetched data.
- Capability inventory: The skill implements network requests and data processing, which are the necessary capabilities for this attack vector.
- Sanitization: The examples demonstrate how to retrieve data but do not provide patterns for sanitizing or filtering the content before the agent acts upon it.
Audit Metadata