skills/adamos486/skills/song-lyrics/Gen Agent Trust Hub

song-lyrics

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (LOW): The skill's core function involves gathering external songwriting insights, which introduces a vulnerability surface if the sourced data contains hidden instructions.
  • Ingestion points: The skill instructions in SKILL.md require the agent to 'research and gather insights' from external sources.
  • Boundary markers: No delimiters or instructions are provided to distinguish external content from the agent's internal logic.
  • Capability inventory: The skill is restricted to presenting information and lacks high-risk capabilities such as file modification, command execution, or network exfiltration.
  • Sanitization: The instructions do not specify any validation or sanitization of the retrieved external content.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 12:14 AM