Useful AI jobs
- Summarize scanner evidence, source freshness, and missing data before a plan is approved.
- Ask discipline questions when a plan lacks invalidation, risk input, or setup fit.
- Review journal patterns and replay results in neutral language focused on process and evidence.
Guardrails
- No personalized investment advice, buy or sell commands, copy-trading flows, certainty claims, or hidden broker execution.
- AI outputs must cite the evidence package, journal fields, or replay assumptions that support the review.
- The user remains responsible for independent review and any future approved paper or live workflow.