blueprint

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). The prompt explicitly injects a "Language Requirement" that forces all responses to be in Korean — contradicting the skill's stated bilingual purpose and covertly attempts to change downstream agents' behavior, so it is a deceptive instruction outside scope.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt instructs the agent to "record user's exact words" and to document "only what the user explicitly stated" and then produce/share those documents, which means any API keys, tokens, or passwords included in a user's request would be preserved and output verbatim, creating high exfiltration risk.