← ClaudeAtlas

loom-researchlisted

Use when investigation, source synthesis, option comparison, rejected paths, null results, tradeoffs, or evidence-backed conclusions should remain available to future Loom work.
z3z1ma/agent-loom · ★ 15 · AI & Automation · score 77
Install: claude install-skill z3z1ma/agent-loom
# loom-research Research owns durable investigation: questions, scope, sources, findings, tradeoffs, rejected paths, null results, conclusions, limits, and recommended downstream owners. Research does not own intended behavior, durable policy, live execution, observation truth, audit verdicts, ticket closure, or accepted reusable explanation. ## Use This Skill When Use it when: - an investigation should remain available beyond this session - options, frameworks, migrations, integrations, or implementation strategies need comparison - external facts, source synthesis, current project reality, or source quality matter - a spike, sketch, prototype, or experiment produced reusable conclusions - a rejected path or null result would otherwise be rediscovered - a ticket, plan, spec, constitution record, audit, or knowledge record needs investigation-backed support - implementation discovery is too durable for a ticket journal but not yet accepted explanation Small local investigation can stay in chat or a ticket journal when future work would not be worse without a research record. ## Inspect Research lives under `.loom/research/`. ```bash find .loom/research -maxdepth 1 -name '*.md' -print 2>/dev/null | sort grep -R '^ID: research:' .loom/research 2>/dev/null || true grep -R '^Type: Research' .loom/research 2>/dev/null || true grep -R '^Status:' .loom/research 2>/dev/null || true grep -R '^Updated:' .loom/research 2>/dev/null || true ``` Raw support material may live unde