agent-workflow
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill identifies a surface for indirect prompt injection where untrusted data enters the agent context.
- Ingestion points: User-provided
messagein the FastAPI/chatand/chat/streamendpoints and CLI input loop inassets/graph-template.py. - Boundary markers: Absent. The user input is directly concatenated or passed into the message list without delimiters like XML tags or explicit 'ignore embedded instructions' warnings.
- Capability inventory: The agent can invoke tools (
search_knowledge_base,get_current_time) and maintain conversation state, though current tools are mocked or benign. - Sanitization: Absent. No escaping or validation is performed on the input string before it is passed to the LLM.
- [SAFE] (SAFE): No hardcoded credentials, malicious network exfiltration, or obfuscated code patterns were found in the script.
Audit Metadata