fixing-metadata
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of instructional guidelines for auditing and fixing HTML metadata such as page titles, descriptions, and Open Graph tags.
- [PROMPT_INJECTION]: No patterns of prompt injection, instruction overrides, or safety filter bypasses were identified.
- [DATA_EXFILTRATION]: No patterns involving the access or transmission of sensitive local data, environment variables, or hardcoded credentials were observed.
- [REMOTE_CODE_EXECUTION]: The skill does not contain instructions to download or execute external scripts, nor does it list any third-party dependencies.
- [COMMAND_EXECUTION]: No dangerous shell commands, privilege escalation attempts, or persistence mechanisms were detected.
- [SAFE]: Regarding indirect prompt injection, the skill includes defensive instructions to escape and sanitize any dynamic or user-generated strings before implementation.
Audit Metadata