Comix Inc. Launches Security Setting Support for Claude Code to Enable Safe Usage

Comix Inc. (Headquarters: Shibuya-ku, Tokyo; Representative Director: Akihiro Suzuki) has launched a "Security Setting Support Plan" for Claude Code and released a "Claude Code Permission Setting Complete Guide" document. This addresses the challenges of safely and continuously operating generative AI in business, particularly for development and business improvement. The initiative responds to the growing concerns among Japanese companies regarding information leakage and dangerous command execution when using generative AI, despite high adoption rates. Statistics from the Ministry of Internal Affairs and Communications' Information and Communications White Papers (2024 and 2025 editions) indicate that while generative AI utilization is increasing (49.7% have policies, 55.2% use it in some operations), approximately 70% of companies are concerned about information leakage risks. An OpenText and Ponemon Institute survey also found that 79% of companies have not reached "AI maturity" in terms of risk assessment. The "Claude Code Permission Setting Complete Guide" covers permission setting importance, a 3-layer scope management model (Managed/Project/User), a 3-layer defense model (Sandbox/Permissions/Hooks), deny/ask/allow evaluation order, and setting examples. The "Security Setting Support Plan" assists in designing initial deny/ask/allow policies, preventing dangerous operations and confidential information access, and establishing continuous operational frameworks. Comix Inc. was established in September 2007 and specializes in generative AI utilization productivity improvement support, B2B sales support, and freelance professional talent introduction.