admin-offboarding
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze various document types, including emails, chat logs, and technical reports provided by the user to assess offboarding risks. While processing untrusted data creates an entry point for potential instructions embedded in that data, the skill does not define or utilize any autonomous tool-execution capabilities (such as shell access or web requests) that could be maliciously triggered by such content. The risk is limited to the agent potentially misinterpreting the analysis due to malicious content within the provided documents.
- Ingestion points: Documents, tables, contracts, tickets, logs, emails, and chat history provided by the user in the context of the offboarding review.
- Boundary markers: None explicitly defined in the prompt to separate user data from instructions.
- Capability inventory: The skill is purely analytical and does not contain subprocess calls, file writes, or network operations.
- Sanitization: No specific filtering or sanitization of external content is mentioned.
Audit Metadata