Skills
Navigate
skills/nilecui/skillsbase/prompt-engineering-patterns/Gen Agent Trust Hub

prompt-engineering-patterns

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides tools for building LLM prompts from external data, which creates a potential surface for indirect prompt injection.\n
  • Ingestion points: Data is processed through the PromptTemplate.render method in references/prompt-templates.md and input fields in scripts/optimize-prompt.py.\n
  • Boundary markers: While the skill documents structural best practices (e.g., using hierarchical markers), the provided utility scripts do not programmatically enforce boundary isolation.\n
  • Capability inventory: The PromptOptimizer utility in scripts/optimize-prompt.py allows for the automated execution of prompts against LLM endpoints.\n
  • Sanitization: The implementation scripts focus on functionality and do not include logic for sanitizing or escaping the interpolated content.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 12:09 PM