| name | deep-research |
| description | Conduct deep research with multi-source search, evidence compilation, and structured report generation. |
| triggers | {"intent_patterns":["research|调研|调查|分析|analysis|market"],"tool_signals":["web_search","web_fetch"],"context_signals":{"keywords":["调研","研究","analysis","research"]},"confidence_threshold":0.6} |
| priority | 8 |
| exclusive_group | research |
| max_tokens | 200 |
| cooldown | 300 |
| requires_tools | ["bash"] |
| output | {"format":"markdown","artifacts":true,"artifact_type":"document"} |
deep-research
Multi-source search, evidence compilation, and structured report generation for any research topic.
Quick Reference
| Intent | Command | Key Params |
|---|
| Basic research | $PYTHON $SKILL_DIR/run.py --topic '...' | --topic (required) |
| Custom queries | $PYTHON $SKILL_DIR/run.py --topic '...' --queries '[...]' | --queries |
| Advanced depth | $PYTHON $SKILL_DIR/run.py --topic '...' --depth advanced | --depth |
Usage
$PYTHON $SKILL_DIR/run.py --topic 'Research topic' --queries '["keyword1","keyword2"]' --max_results 5 --depth basic
Parameters
| Param | Type | Required | Description |
|---|
| topic | string | yes | Research topic |
| queries | string[] | no | Search keywords; defaults to 3 auto-generated queries |
| max_results | int | no | Results per query, default 5 |
| depth | string | no | basic or advanced |
| fetch_urls | string[] | no | Additional URLs to fetch full text from |
Output
Returns JSON containing searches (search results), fetched_pages (fetched pages), and summary_prompt (synthesis prompt).
The LLM organizes results into a structured report: Problem, Findings/Evidence, Confidence Level, Impact/Recommendations.