agent-orchestration

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The content instructs orchestrating code changes by delegating to a CLI worker (cursor-agent) and explicitly recommends/uses auto-approve flags (--force/--trust) and model-override avoidance, which weakens execution controls and could enable automated state-changing actions without interactive confirmation, but it is documentation for tooling rather than clearly malicious.