mantis-architecture

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • Workspace-Local Operations: The skill focuses on analyzing codebase structure and internal learning logs (learnings.jsonl). All file operations are restricted to the local workspace and the intended knowledge base directory (workspace/kb/).
  • Data Synthesis Surface: The skill processes existing workspace data to create documentation. While it acts as a synthesizer for potentially untrusted source code, it does not include features that would facilitate external exfiltration or unauthorized command execution.
  • State Management: The instruction to clear the learnings.jsonl file is a standard method for managing the agent's input queue and preventing redundant processing of the same insights.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 01:21 PM
Security Audit — agent-trust-hub — mantis-architecture