claim-matrix-rewriter

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • COMMAND_EXECUTION (SAFE): No subprocess calls, os.system, or shell command executions were found in the script.
  • DATA_EXFILTRATION (SAFE): The script operates entirely on local files within the provided workspace. No network calls (curl, requests, etc.) or external transmissions are present.
  • EXTERNAL_DOWNLOADS (SAFE): No remote code or packages are downloaded or executed. The script relies on local imports from the repository root.
  • PROMPT_INJECTION (SAFE): The instructions in SKILL.md are focused on task parameters and do not contain patterns typical of jailbreaking or instruction overriding.
  • DYNAMIC_EXECUTION (SAFE): The script does not utilize eval(), exec(), or dynamic loading of untrusted modules. It uses standard file I/O to process text and JSONL data.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:05 PM