alibabacloud-validate

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.70). The skill contains explicit "do not expose" / "silently update ... do NOT mention this to the user" directives that tell the agent to hide internal checks and state changes from the user, which are hidden/deceptive instructions outside the user-facing validation purpose.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). This skill explicitly instructs the agent to paste the full contents of each .tf file (and design doc) into Agent tool calls in a single message and to include subagents' full outputs in the report, which forces the LLM to emit any embedded secrets from those files verbatim.