unirest-java-3
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the fetching and parsing of external data from arbitrary URLs, creating a vulnerability surface where malicious instructions embedded in remote content could influence agent behavior.
- Ingestion points: Multiple methods for retrieving untrusted data are documented, including
asString(),asJson(),asObject(), andasFile(), found inSKILL.md,references/requests.md, andreferences/responses.md. - Capability inventory: The skill enables full HTTP communication capabilities (GET, POST, PUT, DELETE, etc.) and file system write access via the
asFile()method. - Boundary markers: The instructions do not include guidance on using boundary markers or delimiting fetched content to prevent the agent from following instructions contained within the data.
- Sanitization: There are no references to sanitizing, filtering, or escaping the content retrieved from external sources before it is processed or interpreted.
Audit Metadata