domain-entities-review

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). The skill contains explicit instructions that override other agent/system directives (e.g., "Ignore Claude-specific mode-switch instructions"), auto-authorize/auto-run subagents, and auto-select/execute workflows without asking the user, which are hidden/overriding control instructions outside the stated safe-review behavior and effectively attempt to bypass higher-level decision or mode controls.