social-spotify-card
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill treats the user's message as a design 'brief', which is then passed to an external tool to generate file-editing instructions.
- Ingestion points: The user's message is captured as the primary input in
SKILL.md. - Boundary markers: None are defined to isolate the user brief from the agent's internal instructions.
- Capability inventory: The agent uses file-authoring tools and calls
register_open_design_artifactbased on the tool's output. - Sanitization: No explicit sanitization of the user input is described before it is passed to the design tools.
- [SAFE]: The skill configuration in
agents/openai.yamlexplicitly disables implicit invocation, requiring the user to intentionally trigger the skill, which reduces the risk of accidental execution. - [SAFE]: No obfuscated content, hardcoded credentials, or suspicious network operations were identified in the skill files.
Audit Metadata