prompt-engineering
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of a markdown documentation file. It does not include any scripts, executable binaries, or configuration files that could perform actions on a system.
- [PROMPT_INJECTION]: The skill provides examples and templates for prompt engineering but includes specific guidance on mitigating prompt injection. The 'Anti-Patterns' section explicitly recommends using delimiters like
<user_input>to isolate raw input from instructions. - [DATA_EXFILTRATION]: No network-capable commands or functions (such as curl, wget, or fetch) are present in the skill content.
- [CREDENTIALS_UNSAFE]: The document does not contain any hardcoded API keys, tokens, or sensitive credentials. It uses generic placeholders like
{user_query}for demonstration purposes.
Audit Metadata