workflow-write-integration-test

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). The skill contains explicit instructions that override or ignore higher-level/system instructions and to auto-execute/authorize sub-agents and slash-commands (e.g., "Ignore Claude-specific mode-switch instructions", auto-execute slash skill/workflow commands, and "authorize" spawn_agent), which are deceptive or out-of-scope for a test-writing workflow and can override intended agent constraints.