conversation-ng
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill focus is on linguistic and social nuance for conversational modeling.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (chat histories) which is a standard attack surface for this type of assistant. The skill instructions encourage reading full available chat history, which is where external instructions could be embedded. However, it does not possess high-risk capabilities like shell execution or external network access to facilitate a high-severity exploit.
- Ingestion points: Reads chat history from external files via the
Readtool. - Boundary markers: None identified; it processes raw text from chat platforms.
- Capability inventory: Limited to
Read,Write,Edit, andAskUserQuestion. No command execution or network capabilities. - Sanitization: Not present in the instruction set.
Audit Metadata