Vague Triggers
Medium
- Confidence
- 83% confidence
- Finding
- The skill’s activation and action examples are broad and conversationally phrased, without clear trigger boundaries, confirmation requirements, or separation between ordinary dialogue and state-changing wiki commands. In an agent setting, this can cause unintended writes, updates, or linking from ambiguous user text or quoted content, increasing the risk of prompt/command confusion and unauthorized knowledge-base modification.
