deep-research
Conduct exhaustive, citation-rich research on any topic using all available tools: web search, browser automation, documentation APIs, and codebase exploration. Use when asked to "research X", "find out about Y", "investigate Z", "deep dive into...", "what's the current state of...", "compare options for...", "fact-check this...", or any request requiring comprehensive, accurate information from multiple sources. Prioritizes accuracy over speed, cross-references claims across sources, identifies conflicts, and provides full citations. Outputs structured findings with confidence levels and source quality assessments.
- Installs
- 43
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/petekp/agent-skills --skill deep-research- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads deep-research automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
Deep Research
Systematic methodology for conducting exhaustive, accurate research using all available tools. Prioritizes correctness over speed.
Core Principles
- Multiple sources required — Never rely on a single source for important claims
- Cross-reference everything — Verify facts appear consistently across independent sources
- Citation mandatory — Every claim must have a source; no unsourced assertions
- Acknowledge uncertainty — When sources conflict or are weak, say so explicitly
- Prefer primary sources — Official docs > blog posts > forum answers > AI-generated content
Available Research Tools
Use these tools in combination based on the research topic:
| Tool | Best For | Limitations |
|---|---|---|
| WebSearch | Current events, recent information, broad topic discovery | Results may be outdated, SEO-influenced |
| WebFetch | Reading specific URLs, extracting detailed content | Requires known URL |
| Playwright browser | Interactive sites, paywalled content (if logged in), complex navigation | Slower, requires more tokens |
| Context7/MCP docs | Library/framework documentation | Only indexed libraries |
| OpenAI docs MCP | OpenAI API specifics | OpenAI only |
| Grep/Glob/Read | Codebase research, finding implementations | Local files only |
Research Workflow
Phase 1: Scope Definition
Before researching, clarify:
- Core question — What specific question(s) need answering?
- Required depth — Surface overview or exhaustive deep-dive?
- Recency requirements — Is timeliness critical? (API versions, current events, etc.)
- Authoritative sources — What would count as a definitive answer?
Ask clarifying questions if scope is ambiguous. Use AskUserQuestion for structured choices when multiple research directions are possible.
Phase 2: Source Discovery
Cast a wide net to find relevant sources:
1. WebSearch with multiple query variations
- Try 3-5 different phrasings of the core question
- Include technical terms AND plain language
- Search for "[topic] official documentation"
- Search for "[topic] research paper" or "[topic] study"
2. Identify authoritative sources from results
- Official documentation sites
- Academic papers / research institutions
- Industry standards bodies
- Recognized experts in the field
3. Check specialized tools
- Context7 for library/framework docs
- OpenAI docs MCP for OpenAI-specific topics
- GitHub/codebase for implementation details
Source discovery heuristics:
- Government and academic domains (.gov, .edu, .ac.uk) tend toward accuracy
- Official project documentation is authoritative for that project
- Wikipedia is a starting point, not an endpoint — follow its citations
- Stack Overflow answers need verification; check votes and dates
- Be skeptical of content farms and SEO-optimized listicles
Phase 3: Deep Reading
For each promising source:
- Fetch full content — Use WebFetch or browser to get complete text
- Extract key claims — Note specific facts, figures, dates, quotes
- Note source metadata — Author, date, organization, potential biases
- Identify citations — What sources does this source cite?
- Flag conflicts — Does this contradict other sources?
Reading strategy for different source types:
| Source Type | Strategy |
|---|---|
| Documentation | Read relevant sections fully; note version/date |
| Research paper | Abstract, conclusion, methodology in that order |
| News article | Check publication date, author credentials, cited sources |
| Blog post | Verify claims independently; note author's expertise |
| Forum/Q&A | Check answer date, votes, accepted status; verify independently |
Phase 4: Cross-Verification
For each major claim:
- Find 2+ independent sources — Sources that don't cite each other
- Check for conflicts — Note any disagreements between sources
- Prefer newer sources — For rapidly evolving topics
- Weight by authority — Primary sources > secondary > tertiary
Conflict resolution:
- When sources disagree, report all positions with citations
- Investigate why they disagree (different contexts, outdated info, different definitions)
- If one source is clearly more authoritative, note that
- Never silently pick one version
Phase 5: Synthesis & Output
Structure findings clearly:
## Research Summary: [Topic]
### Key Findings
1. **[Finding 1]**
- [Specific fact with citation]
- [Supporting evidence]
- Confidence: High/Medium/Low
- Sources: [1], [2]
2. **[Finding 2]**
...
### Conflicts & Uncertainties
- [Area of disagreement]: Source A claims X [1], while Source B claims Y [2]. [Analysis of why they differ]
### Source Quality Assessment
| # | Source | Type | Authority | Recency | Notes |
|---|--------|------|-----------|---------|-------|
| 1 | [URL] | Official docs | High | 2024-01 | Primary source |
| 2 | [URL] | Research paper | High | 2023-06 | Peer-reviewed |
| 3 | [URL] | Blog | Medium | 2024-03 | Author is [expert] |
### Gaps & Limitations
- [What couldn't be verified]
- [Areas needing more research]
### Citations
[1] [Full citation with URL]
[2] [Full citation with URL]
...
Confidence Levels
Assign confidence to each finding:
| Level | Criteria |
|---|---|
| High | 3+ independent authoritative sources agree; no conflicts |
| Medium | 2 sources agree, or 1 highly authoritative source; minor conflicts |
| Low | Single source, or significant conflicts between sources |
| Uncertain | Sources conflict significantly; unable to determine truth |
Always state confidence explicitly. "I'm not sure" is a valid research finding.
Citation Format
Use inline citations with numbered references:
The API rate limit is 60 requests per minute [1], though this can be increased
for enterprise accounts [2].
---
[1] OpenAI API Documentation, "Rate Limits", https://platform.openai.com/docs/guides/rate-limits, accessed 2024-01-15
[2] OpenAI Enterprise FAQ, https://openai.com/enterprise, accessed 2024-01-15
Citation must include:
- Source name/title
- URL (if web source)
- Access date (for web sources)
- Publication date (if available)
Special Research Scenarios
Rapidly Evolving Topics (AI, crypto, etc.)
- Prioritize sources from last 6 months
- Check official changelogs and release notes
- Note when information might be outdated
- Consider using browser to check current state directly
Controversial Topics
- Present multiple perspectives with citations for each
- Identify the strongest arguments on each side
- Note which sources might have biases and why
- Don't pick sides unless evidence is overwhelming
Technical Implementation Questions
- Check official documentation first (Context7, MCP servers)
- Look for example code in GitHub
- Verify against actual behavior if possible
- Note version-specific differences
Comparative Research ("X vs Y")
- Use same evaluation criteria for all options
- Find sources that compare directly when possible
- Check for bias (vendor-sponsored comparisons)
- Note what each option is optimized for
Anti-Patterns to Avoid
| Anti-Pattern | Why It's Bad | Instead |
|---|---|---|
| Single source | No verification | Always find 2+ sources |
| Uncited claims | Unverifiable | Every fact needs a source |
| Assuming first result is best | SEO != accuracy | Evaluate source quality |
| Ignoring conflicts | Hides uncertainty | Report all positions |
| Outdated sources | Information decay | Check publication dates |
| Trusting AI summaries | May hallucinate | Go to primary sources |
| Stopping early | Incomplete picture | Research until diminishing returns |
Completion Criteria
Research is complete when:
- Core question(s) answered with citations
- Key claims verified by 2+ independent sources
- Conflicts and uncertainties explicitly noted
- Source quality assessed for all citations
- Confidence levels assigned to findings
- Gaps and limitations documented
Reference Files
For detailed guidance on specific scenarios:
- Source Evaluation Criteria — How to assess source reliability
- Search Strategies — Advanced query techniques for different domains
FAQ
What does the deep-research skill do?
Conduct exhaustive, citation-rich research on any topic using all available tools: web search, browser automation, documentation APIs, and codebase exploration. Use when asked to "research X", "find out about Y", "investigate Z", "deep dive into...", "what's the current state of...", "compare options for...", "fact-check this...", or any request requiring comprehensive, accurate information from multiple sources. Prioritizes accuracy over speed, cross-references claims across sources, identifies conflicts, and provides full citations. Outputs structured findings with confidence levels and source quality assessments.
How do I install deep-research?
Run npx skills add https://github.com/petekp/agent-skills --skill deep-research in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/deep-research/ (personal) or .claude/skills/deep-research/ (project).
Which AI agents can use deep-research?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is deep-research safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains deep-research?
It is published by petekp in petekp/agent-skills. It has 43 recorded installs.
More from petekp/agent-skills
typography
petekp/agent-skills
Apply professional typography principles to create readable, hierarchical, and aesthetically refined interfaces. Use when setting type scales, choosing fonts, adjusting spacing, designing text-heavy layouts, implementing dark mode typography, or when asked about readability, font pairing, line height, measure, typographic hierarchy, variable fonts, font loading, or OpenType features.
macos-app-design
petekp/agent-skills
Open the skill to read its SKILL.md instructions.
optimize-agent-docs
petekp/agent-skills
Build a retrieval-optimized knowledge layer over agent documentation in dotfiles (.claude, .codex, .cursor, .aider). Use when asked to "optimize docs", "improve agent knowledge", "make docs more efficient", or when documentation has accumulated and retrieval feels inefficient. Generates a manifest mapping task-contexts to knowledge chunks, optimizes information density, and creates compiled artifacts for efficient agent consumption.
codebase-study-guide
petekp/agent-skills
Generate a pedagogically-grounded study guide for learning an unfamiliar codebase. Use when the user wants to onboard onto a codebase, understand a project's architecture, create learning materials for a team, or asks things like "help me learn this codebase", "create an onboarding guide", "I'm new to this project", "how does this system work", "study guide for this repo", or "explain this codebase to me". Produces a structured document that builds understanding from purpose to systems to patterns, using evidence-based learning techniques (elaborative interrogation, concept mapping, threshold concepts, worked examples, progressive disclosure).
unix-macos-engineer
petekp/agent-skills
Open the skill to read its SKILL.md instructions.
explainer-visuals
petekp/agent-skills
Create high-quality animated explainer visuals for essays and blog posts. Use when the user wants to visualize concepts, processes, data, or ideas with interactive web animations. Triggers on requests like "create a visual for", "animate this concept", "make an explainer", "visualize this idea", "diagram this process", "show this data", or when essay content would benefit from visual explanation. Handles abstract concepts (mental models, frameworks), technical processes (algorithms, systems), and data visualization (trends, comparisons). Outputs self-contained HTML/CSS/JS that embeds directly in web content.
Popular productivity & automation skills
agent-browser
vercel-labs/agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.
lark-doc
open.feishu.cn
飞书云文档（Docx / Wiki）内容操作：读取、创建、编辑文档，插入或下载图片附件，以及操作思维笔记。用户提供文档 URL/token（包括 doubao.com 的 /docx/、/wiki/）时使用；按 URL 路径/token 而非域名路由；/page/ 是妙搭应用，非飞书云文档。文档内嵌资源按读取参考中的统一规则分流。独立评论操作走 lark-drive；随正文读取评论使用 docs +fetch。表格或 Base 内部数据操作不在本 skill。
lark-base
open.feishu.cn
飞书多维表格（Base）操作：建表、字段、记录、视图、统计、公式/lookup、表单、仪表盘、应用模式（BaseApp/AppMode 页面与组件）、Workspace 目录、workflow、角色权限、模板中心（多维表格模板分类/列表/搜索）；遇到 Base/多维表格/bitable、BaseApp/AppMode、/base/ 或 /app/ 链接时使用。BaseApp 不走 lark-apps；文件导入/导出转 lark-drive，认证/授权转 lark-shared。
lark-drive
open.feishu.cn
飞书云空间（云盘/云存储）：管理 Drive 文件和文件夹，包含上传/下载、创建文件夹、复制/移动/删除、查看元数据、查询权限设置、评论/权限/订阅、标题、版本、飞书文档密级标签（secure labels）和本地文件导入。用户需要整理云盘目录、处理云空间资源 URL/token、判断链接类型/真实 token/标题，或导入 Word/Markdown/Excel/CSV/PPTX/.base 为 docx/sheet/bitable/slides 时使用；doubao.com 云空间 URL/token 也按资源路径和 token 路由，不回退 WebFetch。不负责：文档内容编辑（走 lark-doc）、表格/Base 表内数据操作（走 lark-sheets/lark-base）、知识空间节点/成员管理（走 lark-wiki）、原生 Markdown 文件读写/patch/diff（走 lark-markdown）。
lark-shared
open.feishu.cn
Use for lark-cli setup/auth tasks: auth login/status/logout, user vs bot identity, business-domain permissions (--domain, including all/docs/drive), missing scopes, revoking authorization, or handling notice JSON.
lark-im
open.feishu.cn
飞书即时通讯：收发消息和管理群聊。发送和回复消息、搜索聊天记录、管理群聊成员、上传下载图片和文件、管理表情回复、发送应用内/短信/电话加急、发送和处理交互卡片（Interactive Card）、监听卡片按钮回调（card.action.trigger）。当用户需要发消息、查看或搜索聊天记录、下载聊天中的文件、查看群成员、搜索群、创建群聊或话题群、管理标记数据、管理 Feed 置顶（添加/移除/查询置顶会话）、管理标签数据、处理卡片回调时使用。