Record research lessons with AI
Keep decisions, experiments and reusable lessons connected across your research projects.
Record research lessons with AI so an experiment's result and the reasoning behind a choice survive the session that produced them. research-log maintains local project goals, state and decision journals, then separates specific events from lessons that might transfer to other work.
Use it to register a project, save session context, recall an earlier solution or review progress across projects. A recurring lesson can become a proposed rule, with your approval. Such rules should surface relevant evidence before a choice; they do not replace your judgement or turn one unexplained failure into a universal principle.
What it does
- Connect decisions to their original goal and evidence
- Capture the outcome and cause of an experiment
- Recover a previous solution when a blocker returns
- Save current context for the next research session
- Review drift from each project's current focus
- Promote recurring lessons through an approval step
Install and use
- Use a file-capable agent with authorised project paths; QMD is needed for indexed recall and review.
- Register the project and define its goal and current focus.
- Supply the decision or result, separating established causes from uncertain interpretations.
- Review proposed rule and archive changes, and check that the search index refreshed when used.
FAQ
Source
Made by Tae-Geun Kim. Original skill: research-log, published on GitHub under the MIT licence.
View on GitHub