Description-Behavior Mismatch
Medium
- Confidence
- 92% confidence
- Finding
- The skill is presented as a trading AI, but it also directs the agent to perform content marketing and lead-generation tasks such as writing Substack posts and Twitter/DM copy. This scope expansion is risky because it enables unrelated outward-facing actions that can spam, manipulate users, or cause brand/reputation harm without clear user consent or boundary controls.
