context
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes historical session data that may contain untrusted content from previous interactions. \n
- Ingestion points: Data is retrieved from the
get_context_primertool and subsequent calls toget_session_digest, including fields likeclosing.userandclosing.assistant(SKILL.md). \n - Boundary markers: The instructions lack explicit delimiting or warnings to ignore instructions found within the historical data. \n
- Capability inventory: The skill uses MCP tools including
get_context_primer,search_sessions, andget_session_digest. \n - Sanitization: No sanitization, filtering, or validation is applied to the historical data before it is synthesized into a prose brief.
Audit Metadata