This is the execution layer for individual research dimensions in a structured investigation workflow. It reads a dimension from plan.json, runs multi-round searches across source types (academic papers via sn-search-academic, code via sn-search-code, Chinese or English social platforms, official documents via web search), cross-verifies claims, and writes a sub-report with evidence tables to sub_reports/{dimension_id}.md. The core model is three-pass: breadth discovery with general search, targeted deep dives with specialized skills where available, and verification at original sources. It won't write the final report or replan your dimensions, just execute one dimension thoroughly and update its status to done. Use it when your plan is locked and you need one slice researched properly with traceable evidence.
npx -y skills add opensensenova/sensenova-skills --skill sn-dimension-research --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
sickn33/antigravity-awesome-skills