prompt-engineering

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOW
Full Analysis
  • [PROMPT_INJECTION] (SAFE): The skill provides instructions on how to engineer prompts for other tasks but does not contain commands to override the host agent's safety protocols or system instructions.
  • [DATA_EXFILTRATION] (SAFE): There are no file system access commands or network requests.
  • [EXTERNAL_DOWNLOADS] (SAFE): No external dependencies or remote scripts are referenced for execution.
  • [COMMAND_EXECUTION] (SAFE): The skill consists of documentation and code examples for prompt templating. It does not execute arbitrary shell commands.
  • [INDIRECT_PROMPT_INJECTION] (LOW): While the skill discusses methods for interpolating data into prompts (e.g., {user_code}, {retrieved_context}), these are provided as educational patterns. The skill itself does not ingest untrusted data. Developers implementing these patterns should apply proper boundary markers and sanitization as best practices.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 04:52 AM