Vague Triggers
Medium
- Confidence
- 94% confidence
- Finding
- The skill description is broadly scoped ('mirror frontend and backend code across the stack' and 'scaffold full-stack features from one side'), which can cause the agent to activate on many ordinary coding prompts that are not actually requesting this capability. Over-broad activation increases the chance the skill reads unrelated workspace files or generates code changes in contexts the user did not intend, expanding attack surface and enabling prompt-routing abuse.
