knowledge-handoff
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from external conversations and chat logs to create permanent knowledge artifacts on the file system.\n
- Ingestion points: Data is captured from sources such as Codex chats, external conversations, and various source materials as described in SKILL.md.\n
- Boundary markers: No specific delimiters or boundary markers are defined in the instructions to isolate untrusted content from the system prompt.\n
- Capability inventory: The skill performs file system writes to paths including
raw/,wiki/, and.agents/logs/as part of its core logic (SKILL.md).\n - Sanitization: The skill includes a mandatory instruction to "Drop: secrets, credentials, and unnecessary personal data" before committing any information to the knowledge base (SKILL.md).
Audit Metadata