distill-response
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process text from previous assistant responses, which constitutes an interaction with untrusted data. While this provides a theoretical surface for malicious instructions to influence the distillation logic, the risk is negligible because the skill possesses no dangerous capabilities such as tool execution or network access.\n
- Ingestion points: The skill uses content from preceding assistant responses as its primary input material, as defined in the
Route the Requestsection ofSKILL.md.\n - Boundary markers: The instructions lack specific delimiting markers to encapsulate the source material and prevent it from being interpreted as instructions by the underlying model.\n
- Capability inventory: The skill is restricted to text generation; analysis of all files confirms no subprocess spawning, network requests, or file-system write operations are present.\n
- Sanitization: There are no explicit steps for sanitizing or filtering the content of the processed responses.
Audit Metadata