research-deep

1# Research Deep - Deep Research
2 
3## Trigger
4`/research-deep`
5 
6## Workflow
7 
8### Step 1: Auto-locate Outline
9Find `*/outline.yaml` file in current working directory, read items list, execution config (including items_per_agent).
10 
11### Step 2: Resume Check
12- Check completed JSON files in output_dir
13- Skip completed items
14 
15### Step 3: Batch Execution
16- Batch by batch_size (need user approval before next batch)
17- Each agent handles items_per_agent items
18- Launch web-search-agent (background parallel, disable task output)
19 
20**Parameter Retrieval**:
21- `{topic}`: topic field from outline.yaml
22- `{item_name}`: item's name field
23- `{item_related_info}`: item's complete yaml content (name + category + description etc.)
24- `{output_dir}`: execution.output_dir from outline.yaml (default: ./results)
25- `{fields_path}`: absolute path to {topic}/fields.yaml
26- `{output_path}`: absolute path to {output_dir}/{item_name_slug}.json (slugify item_name: replace spaces with _, remove special chars)
27 
28**Hard Constraint**: The following prompt must be strictly reproduced, only replacing variables in {xxx}, do not modify structure or wording.
29 
30**Prompt Template**:
31```python
32prompt = f"""## Task
33Research {item_related_info}, output structured JSON to {output_path}
34 
35## Field Definitions
36Read {fields_path} to get all field definitions
37 
38## Output Requirements
391. Output JSON according to fields defined in fields.yaml
402. Mark uncertain field values with [uncertain]
41…