adding-animations

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [Indirect Prompt Injection] (HIGH): The skill has a high vulnerability surface for Category 8 attacks as it explores potentially untrusted content with high-privilege tools. 1. Ingestion points: The codebase is explored via Read, Glob, and Grep tools (SKILL.md). 2. Boundary markers: None; there are no instructions to ignore or delimit embedded commands within the code. 3. Capability inventory: The skill uses Write, Edit, and Task tools, which allow for file modification and arbitrary command execution (SKILL.md). 4. Sanitization: No content validation or sanitization is mentioned.
  • [Command Execution] (MEDIUM): The 'Task' tool enables shell execution, which poses a significant risk if the agent is misled by malicious comments or strings encountered during codebase exploration.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 09:26 AM