album-art-director
Pass
Audited by Gen Agent Trust Hub on Mar 27, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill uses a vendor-provided tool 'bitwize-music-mcp' from 'bitwize-music-studio' which is consistent with its stated purpose of managing music-related visual concepts.
- [SAFE]: The skill implements a secure configuration override mechanism, allowing users to provide custom visual preferences via a local file ('album-art-preferences.md') without requiring external network access or risky permissions.
- [PROMPT_INJECTION]: The skill processes untrusted data from album concepts and user preferences to construct AI art prompts. This presents a surface for indirect prompt injection where malicious instructions could be embedded in metadata or tracklists. However, the impact is limited to the generated prompt text and does not grant unauthorized system access.
- Ingestion points: Reads project files and arguments for album details and user preferences via the Read tool.
- Boundary markers: No explicit delimiters or boundary instructions are used to separate ingested content from system instructions in the generated prompts.
- Capability inventory: The skill has access to file modification tools ('Edit', 'Write') and the vendor-specific MCP tool.
- Sanitization: External data is interpolated directly into prompts without escaping or validation.
Audit Metadata