video-script
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [DATA_EXFILTRATION]: The skill is instructed to access and read brand-specific information, compliance rules, and standard operating procedures from the user's home directory (e.g., ~/.claude-marketing/brands/). While these paths appear to be application-specific data for the marketing suite, accessing the home directory constitutes a potential data exposure risk.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests external data from JSON profiles and Markdown templates and uses it to guide the script generation process without explicit sanitization.
- Ingestion points: Loads data from ~/.claude-marketing/brands/{slug}/profile.json, _manifest.json, and other local brand/SOP files.
- Boundary markers: Absent. No delimiters or specific instructions are provided to the agent to prevent it from executing instructions found within the loaded files.
- Capability inventory: The agent has the capability to read local files and generate production-ready content, which could be manipulated by content within those files.
- Sanitization: Absent. The skill does not mention any validation or sanitization steps for the data retrieved from the filesystem.
Audit Metadata