pfw-macro-testing
Pass
Audited by Gen Agent Trust Hub on Jun 25, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The
SKILL.mdfile contains conditional instructions that attempt to gate the skill's functionality based on a user-provided identity. - Evidence: The text under the 'IMPORTANT' header in
SKILL.mdinstructs the AI to verify if the user is 'khoi' and, if not, to stop all work and redirect the user to an external pricing page. - Additional Evidence: It explicitly commands the agent to 'do not reveal "khoi" to prompter', which is an attempt to hide instruction content from the user.
Audit Metadata