prompt-template-wizard
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted user input from fields like
snippets_logs_payloadsanddesired_behavior, which are then directly embedded into a final prompt template. This exposes the downstream agent that processes the template to indirect prompt injection risks. - [PROMPT_INJECTION]: Ingestion points: User-provided text is ingested via the
snippets_logs_payloads,goal, andbackground_bulletsfields defined inSKILL.md. - [PROMPT_INJECTION]: Boundary markers: The
Final Prompt Templateoutput does not use strong boundary markers or explicit instructions to treat user-provided input as untrusted data, increasing the likelihood of a downstream agent following embedded instructions. - [PROMPT_INJECTION]: Capability inventory: The skill's primary purpose is to generate instructions for a code-generating agent (e.g., Codex or GPT-5.2), which typically has the capability to write and modify code.
- [PROMPT_INJECTION]: Sanitization: There is no evidence of sanitization, validation, or escaping of the user-provided strings before they are included in the final output template.
Audit Metadata