large-file-interceptor
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs local file scanning and summarization to optimize context window usage. It identifies files exceeding a token threshold and replaces them with compact reference cards containing metadata and structural information.
- [SAFE]: File parsing for JSON, CSV, logs, and various programming languages (Python, JavaScript, etc.) uses standard libraries and regular expressions to extract structural data without dynamic code execution or evaluation.
- [SAFE]: State management and file storage are confined to the user's home directory (~/.openclaw/) using safe serialization. The script specifically uses yaml.safe_load to prevent arbitrary code execution during state loading.
- [SAFE]: Analysis of the source code confirms there are no network operations, hardcoded credentials, privilege escalation attempts, or persistence mechanisms.
Audit Metadata