prompt-engineering-patterns

Pass

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [Prompt Injection] (LOW): The skill's primary function is to wrap user-provided data in prompt templates, creating a surface for indirect prompt injection.\n
  • Ingestion points: Numerous templates in assets/prompt-template-library.md, SKILL.md, and references/system-prompts.md ingest external data via variables like {query}, {text}, and {message}.\n
  • Boundary markers: While some templates utilize labels (e.g., 'Context:', 'Question:'), they lack robust delimiters or specific 'ignore embedded instructions' warnings necessary to prevent adversarial input from hijacking the agent's logic.\n
  • Capability inventory: The scripts/optimize-prompt.py script contains logic to programmatically render and execute these templates via an LLM client, demonstrating the path from untrusted input to model execution.\n
  • Sanitization: There is no evidence of input validation, sanitization, or escaping of placeholder variables in the provided scripts or templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 18, 2026, 06:31 AM