Vague Triggers
Medium
- Confidence
- 95% confidence
- Finding
- The trigger list includes broad, generic terms like "session," "context," and "conversation-state," which are likely to match many ordinary interactions and cause the skill to activate outside its intended scope. Over-broad auto-invocation increases the attack surface by injecting this skill's behavior into unrelated tasks, potentially interfering with agent behavior or causing unintended state retention in contexts where temporary memory handling was not requested.
