From kb
Conduct preliminary research on a target topic and generate a research outline. Used for academic research, benchmark research, technology selection, and similar scenarios.
How this skill is triggered — by the user, by Claude, or both
Slash command
/kb:researchThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
> **Attribution:** Originally authored by [Weizhena](https://github.com/Weizhena/Deep-Research-skills). Included with attribution for use in the Deep query workflow.
Attribution: Originally authored by Weizhena. Included with attribution for use in the Deep query workflow.
/research <topic>
Based on the topic, use the model's existing knowledge to generate:
Output {step1_output} and use AskUserQuestion to confirm:
Use AskUserQuestion to inquire about the time range (e.g., last 6 months, 2024 to present, no limit).
Parameter Collection:
{topic}: Research topic input by user{YYYY-MM-DD}: Current date{step1_output}: Complete output content generated in Step 1{time_range}: Time range specified by userHard Constraint: The following prompt must be strictly reproduced, only replacing variables in {xxx}, no rewriting of structure or wording allowed.
Launch 1 web-search-agent (background), Prompt Template:
prompt = f"""## Task
Research Topic: {topic}
Current Date: {YYYY-MM-DD}
Based on the following preliminary framework, supplement the latest items and recommended research fields.
## Existing Framework
{step1_output}
## Objectives
1. Verify if existing items miss important objects
2. Supplement items based on missing objects
3. Continue searching for {topic}-related items within {time_range} and supplement
4. Supplement new fields
## Output Requirements
Return structured results directly (do not write files):
### Supplemented Items
- item_name: Brief explanation (why should it be included)
...
### Recommended Additional Fields
- field_name: Field description (why this dimension is needed)
...
### Information Sources
- [Source 1](url1)
- [Source 2](url2)
"""
One-shot Example (assuming research on AI Coding development history):
## Task
Research Topic: AI Coding Development History
Current Date: 2025-12-30
Based on the following preliminary framework, supplement the latest items and recommended research fields.
## Existing Framework
### Items List
1. GitHub Copilot: Developed by Microsoft/GitHub, first mainstream AI programming assistant
2. Cursor: AI-first IDE, based on VSCode
...
### Field Framework
- Basic Info: name, release_date, company
- Technical Features: underlying_model, context_window
...
## Objectives
1. Verify if existing items miss important objects
2. Supplement items based on missing objects
3. Continue searching for AI Coding Development History-related items within 2024 to present and supplement
4. Supplement new fields
## Output Requirements
Return structured results directly (do not write files):
### Supplemented Items
- item_name: Brief explanation (why should it be included)
...
### Recommended Additional Fields
- field_name: Field description (why this dimension is needed)
...
### Information Sources
- [Source 1](url1)
- [Source 2](url2)
Use AskUserQuestion to ask if the user has predefined field files, read and merge if available.
Merge {step1_output}, {step2_output} and user's existing fields to generate two files:
outline.yaml (items + configuration):
fields.yaml (field definitions):
./{topic_slug}/outline.yaml and fields.yaml{current_working_directory}/{topic_slug}/
├── outline.yaml # items list + execution configuration
└── fields.yaml # field definitions
/research-add-items - Supplement items/research-add-fields - Supplement fields/research-deep - Start deep researchnpx claudepluginhub rvk7895/llm-knowledge-bases --plugin kbRead research outline and launch independent agents for in-depth research on each item. Task output disabled.
Guides users through building structured research queries with topic, scope, and source quality selection before starting deep research.
Orchestrates multi-step research projects by scaffolding plans, generating atomic tasks with dependencies, and persisting state via Tasks or files. Use for new research or complex workflows.