ai-music-and-sound
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its data processing instructions.
- Ingestion points: The instructions in SKILL.md and references/scope-and-connections.md state that the agent must read external files including the 'brand-profile' and the 'video/asset' being scored.
- Boundary markers: No specific delimiters or instructions are provided to separate these external inputs from the agent's core instructions, making it possible for content in those files to influence the agent's behavior.
- Capability inventory: While the provided files do not include executable scripts, the skill's instructions direct the agent to perform analysis and advisory tasks. In a default environment where tools are not restricted in the frontmatter, the agent may have access to tools that could be abused via injection.
- Sanitization: No input validation, escaping, or filtering of the external content is described.
- [PROMPT_INJECTION]: The skill presents hypothetical legal and industry information (e.g., settlements dated late 2025/2026) as 'verified' facts within a future-dated persona. This metadata poisoning could lead to the agent providing inaccurate information about the current real-world legal status of AI music tools.
- [NO_CODE]: The analyzed skill package consists entirely of markdown-based instructions, frameworks, and JSON evaluation data. No executable code or binary files were found.
Audit Metadata