skills/xmcp-dev/skills/prompt-design/Gen Agent Trust Hub

prompt-design

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill functions by interpolating user-provided data (parameters, code snippets, diffs) directly into prompt handlers and TypeScript templates.
  • Ingestion points: Data enters through {code}, {diff}, {handler}, and other user-defined parameters in src/prompts/*.ts handlers.
  • Boundary markers: Partial boundary markers (triple backticks) are present in the templates, but no explicit 'ignore embedded instructions' warnings are included.
  • Capability inventory: The skill primarily performs file creation in src/prompts/ and suggests running pnpm build via the agent.
  • Sanitization: No evidence of sanitization or escaping of the interpolated strings is found in the provided templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:22 PM