raffle-winner-picker
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from untrusted external sources, which may contain malicious instructions targeting the agent.
- Ingestion points: Untrusted data enters the agent context via external Google Sheet URLs, local spreadsheet files (e.g.,
entries.csv), and user-pasted lists as described inSKILL.md. - Boundary markers: The instructions do not define clear delimiters or provide instructions to the agent to ignore any embedded directives within the source data rows or list items.
- Capability inventory: To perform its tasks, the skill utilizes the agent's file-reading and web-browsing capabilities to fetch participant data from the specified sources.
- Sanitization:
SKILL.mddoes not specify any sanitization, validation, or escaping logic to handle content retrieved from these external sources before processing.
Audit Metadata