mermaid-gantt-chart-generator

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • Indirect Prompt Injection (HIGH): The skill operates within a high-risk capability tier due to its access to the Bash tool while processing untrusted external content. Findings: 1. Ingestion points: User prompts containing Mermaid Gantt chart descriptions (identified in 'When to Use'). 2. Boundary markers: None specified to separate user data from agent instructions. 3. Capability inventory: 'Bash', 'Write', 'Edit', 'Read', and 'Grep' tools. 4. Sanitization: No evidence of input validation or sanitization before tool use.
  • Command Execution (HIGH): The explicit request for 'Bash' tool access is over-privileged for a visual content generator and creates a direct execution vector for injected payloads.
  • Metadata Poisoning (MEDIUM): The skill uses repetitive keyword triggers ('mermaid gantt chart generator') in its description, which is a pattern associated with SEO-style metadata poisoning to manipulate agent skill selection over safer alternatives.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 07:01 AM