curiosity

To surface research topics from vault notes, the curiosity Skill analyzes recent vault activity, daily observations, and goals to identify and prioritize high-value candidate investigations. It scores opportunities by Expected Information Gain to minimize manual effort in discovering research paths.

Expected Information Gain (EIG) is a scoring mechanism used to rank candidate research topics identified from vault activity. By evaluating EIG, the system prioritizes investigations that promise the highest learning value and can auto-trigger research for top-ranked topics.

Yes, you can execute a dry-run to test vault research discovery without auto-triggering learning tasks. The Skill respects configured options for dry-run execution, allowing you to review candidate topics and thresholds before committing to automated research actions.

Configuring thresholds and maximum candidates for vault investigation prioritization is handled through the Skill's configured options. You can set specific EIG score thresholds and limit the maximum number of candidate topics surfaced to control the volume of automated research tasks.

Yes, vault research automation logs provenance for each candidate topic it surfaces. This ensures you can trace exactly which recent vault notes, daily observations, or goals contributed to the identification and prioritization of any specific research opportunity.

To execute research tasks after surfacing high-value vault topics, the Skill integrates directly with the /learn command. It auto-triggers learning for high-EIG topics, executing the research tasks and automatically logging the results back into your system.