technical-video-script
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXFILTRATION]: The skill encourages secure behavior by instructing users to redact or seed sensitive data—specifically customer names, tenant IDs, internal hostnames, and pricing—to ensure they do not appear in final video recordings.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text material such as blog posts, documentation, and pull requests. Although this creates an indirect injection surface, the risk is negligible as the skill operates within a restricted text-generation context without access to sensitive APIs, network operations, or file system writes.
- [REMOTE_CODE_EXECUTION]: The instructions contain references to other specialized skills within the same author's ecosystem (e.g., samber/developer-relations-skills). These are legitimate platform-level navigation aids for the agent and do not represent a remote code execution risk.
Audit Metadata