loom-research

loom-research

Research owns reusable discovery.

Use it when the work would benefit from not having to rediscover the same reasoning later.

What This Skill Owns

• investigations
• option comparisons
• experiments
• spike and sketch conclusions as research variants when discovery should remain citable
• evidence synthesis
• rejected options and null results worth preserving
• deferred questions that are not yet ready for their own investigation
• conclusions and recommendations grounded in evidence

Use This Skill When

• the project lacks evidence for a decision
• multiple options need comparison
• an experiment result should remain citable
• a spike or sketch produced conclusions, rejected variants, or null results that should become reusable research; use loom-spike for procedural spike or sketch workflow detail
• a significant implementation discovery should survive beyond the ticket journal
• a rejected option or null result would otherwise have to be rediscovered
• open questions are accumulating and need a durable home before they are ready to be investigated

Do Not Use This Skill When

• the work is now a behavior contract
• the work is now a rollout strategy
• the work is just live progress tracking

Good Research Questions

A strong research note answers:

• what was being investigated
• why it mattered
• how it was investigated
• what evidence was gathered
• what conclusions are justified
• what downstream work should do next

Done Means

• another agent could reuse the note without rerunning the whole investigation
• uncertainty is visible
• recommendations are bounded by actual evidence

Read In This Order

Read immediately for normal research creation or review:

1. references/research-shape.md when structuring investigation, rejected options, null results, conclusions, or recommendations.
2. references/source-handling.md when external sources, current facts, or source quality matter.
3. templates/research.md only when creating or substantially reshaping a research record.