Vague Triggers
Medium
- Confidence
- 81% confidence
- Finding
- The skill presents very broad activation cues such as asking about light-token, defi, payments, or program migration, which are common topics in normal Solana development discussions. This can cause the agent to auto-invoke the skill in situations where the user did not explicitly request it, increasing the chance of unwanted capability loading, overreach, or routing into external documentation and MCP-backed guidance. In this context, the risk is elevated because the skill is large, installs multiple sub-skills, and encourages automatic use across a wide domain.
