The Agent Skills Directory

[SAFE]: The skill defines a standard workflow for fraud investigation, including assessment, initiation, evidence gathering, and reporting. No malicious commands or instructions were found.
[SAFE]: No remote code execution or external downloads are present. The skill operates entirely within the provided context.
[SAFE]: No data exfiltration or credential theft patterns were detected. It processes investigative data but does not attempt to send it to external domains.
[SAFE]: No obfuscation or prompt injection techniques are used. The text is clear and follows its stated purpose.
[SAFE]: Accessing project-level context files (like CLAUDE.md) is a standard practice for maintaining consistency across tasks and does not constitute a security risk in this context.

fraud-investigation