ai-prompt-engineering
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's content is instructional and defensive. It identifies prompt injection and data exfiltration as primary threats and provides guidance on how to secure applications against them using industry standards like the OWASP LLM Top 10.- [PROMPT_INJECTION]: The skill provides templates in
SKILL.mdthat accept untrusted external data (e.g.,{{input_data}},{{chunks_with_ids}}). While this represents a surface for indirect prompt injection, the skill includes explicit mitigation strategies such as boundary delimiters, schema validation, and instruction/data separation. The inclusion of these guardrails aligns with security best practices for prompt engineering.
Audit Metadata