signal-detector
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by processing untrusted data and persisting it to a knowledge base.
- Ingestion points: Ingests data from every inbound message as specified in SKILL.md.
- Boundary markers: Absent; there are no instructions to use delimiters or warnings to separate captured content from other instructions when stored in brain pages.
- Capability inventory: The skill uses tools such as
put_page,add_link, andadd_timeline_entryto modify persistent storage based on external input. - Sanitization: Absent; instructions explicitly mandate capturing user input using exact phrasing without paraphrasing or filtering.
Audit Metadata