deep-research
Deep research and analysis tool. Generates comprehensive HTML reports on any topic, domain, paper, or technology. Use when user asks to research, analyze, investigate, deep-dive, or generate a report on any subject. Supports academic papers (arXiv), technologies, trends, comparisons, and general topics.
Profile →npx skills add actionbook/actionbook --skill deep-research --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Analyzes a topic, domain, or paper and produces a rich HTML report via Actionbook browser automation and json-ui rendering. It is designed for users who want deep research, analysis, and reporting on topics, academic papers (arXiv), technologies, trends, and comparisons.
How it works
The skill instructs the agent to perform a structured research workflow that includes planning search strategy, querying verified selectors with Actionbook API, performing arXiv Advanced Search using those selectors, supplementing with Google/Bing results, deep-reading sources, and synthesizing findings into a json-ui JSON report. It enforces the MUST USE Actionbook CLI for all web browsing steps, with explicit commands to open pages, interact with forms (selects, checkboxes, input fields), extract text, take snapshots, and close the browser. The workflow includes:
- Step 1: Plan 5-8 search queries across definition, latest developments, technical details, comparisons, expert opinions, and use cases.
- Step 2: Query Actionbook API for verified selectors before browsing, to feed arXiv Advanced Search and other sites.
- Step 3: Use arXiv Advanced Search with the obtained selectors to perform multi-field, multi-criteria searches.
- Step 4: Supplement with Google/Bing searches via Actionbook browser.
- Step 5: For each relevant URL, re-check indexing with Actionbook API, then gather content using verified selectors or page text.
- Step 6: Synthesize findings into an executive overview, key findings, detailed analysis, evidence, implications, and sources.
- Step 7: Generate a json-ui JSON Report at the path ./output/<topic-slug>.json.
- Step 8: Render the JSON report to HTML using one of several render methods, always attempting fallbacks, and ensuring absolute paths.
- Step 9–10: Open the generated report in a browser and close the Actionbook browser.
Parameters
Topic is required; optional flags include --lang and --output. The tool supports topics including arXiv papers, technologies, and general topics, and provides a bilingual output option.
When to use it
Use when the user asks to research, analyze, investigate, deep-dive, or generate a report on any subject. It supports academic papers (arXiv), technologies, trends, comparisons, and general topics.
What it can touch
The skill interacts with the Actionbook CLI to drive web browsing, including commands such as:
- actionbook browser open <url>
- actionbook browser snapshot
- actionbook browser text [selector]
- actionbook browser click <selector>
- actionbook browser close
- npx @actionbookdev/json-ui render /absolute/path/to/report.json -o /absolute/path/to/report.html
- json-ui render /absolute/path/to/report.json -o /absolute/path/to/report.html
- node "$(git rev-parse --show-toplevel)/packages/json-ui/dist/cli.js" render /absolute/path/to/report.json -o /absolute/path/to/report.html
Output
A JSON report following the @actionbookdev/json-ui schema, saved to the path specified by --output (default ./output/<topic-slug>.json), and an HTML rendering of that report.
Caveats
License is Apache-2.0. The workflow requires the Actionbook toolchain and may depend on indexable sites and verified selectors. The guidance emphasizes using Actionbook selectors, and warns against relying solely on generic web fetches. The instruction set enforces absolute paths for render outputs and requires always closing the browser at the end.
# Deep Research Analyze any topic, domain, or paper and generate a beautiful HTML report using Actionbook browser automation and json-ui rendering. ## Usage ``` /deep-research:analyze <topic> /deep-research:analyze <topic> --lang zh /deep-research:analyze <topic> --output ./reports/my-report.json ``` Or simply tell Claude: "帮我深度研究 XXX 并生成报告" / "Research XXX and generate a report" ### Parameters | Parameter | Required | Default | Description | |-----------|----------|---------|-------------| | `topic` | Yes | - | The subject to research (any text) | | `--lang` | No | `both` | Language: `en`, `zh`, or `both` (bilingual) | | `--output` | No | `./output/<topic-slug>.json` | Output path for JSON report | ### Topic Detection | Pattern | Type | Strategy | |---------|------|----------| | `arxiv:XXXX.XXXXX` | Paper | **arXiv Advanced Search** (Step 2b) + ar5iv deep read | | `doi:10.XXX/...` | Paper | Resolve DOI, then **arXiv Advanced Search** for related work | | Academic keywords (paper, research, model, algorithm) | Academic topic | **arXiv Advanced Search** (Step 2b) + Google for non-academic sources | | URL | Specific page | Fetch and analyze the page | | General text | Topic research
- Usage
- Parameters
- Topic Detection
- Architecture
- Why Actionbook, Not WebFetch/WebSearch?
- MUST USE Actionbook CLI
- Complete Workflow
- Step 1: Plan Search Strategy
- Step 2: Query Actionbook API for Selectors (ALWAYS DO THIS FIRST)
- Step 3: arXiv Advanced Search (Using Actionbook Selectors)
- Step 4: Supplement with Google / Bing Search
- Step 5: Deep Read Sources
- Step 6: Synthesize Findings
- Step 7: Generate json-ui JSON Report
actionbook browser open <url> # Navigate to page actionbook browser snapshot # Get accessibility tree actionbook browser text [selector] # Extract text content actionbook browser screenshot [path] # Capture visual actionbook browser click <selector> # Click element actionbook browser close # Close browser (ALWAYS do this at end) Search for indexed actions by domain actionbook search "<keywords>" -d "<domain>" Get detailed selectors for a specific page actionbook get "<domain>:/<path>:<area>"
What does the deep-research skill do?
Deep research and analysis tool. Generates comprehensive HTML reports on any topic, domain, paper, or technology. Use when user asks to research, analyze, investigate, deep-dive, or generate a report on any subject. Supports academic papers (arXiv), technologies, trends, comparisons, and general topics.
How do I install it?
Run `npx skills add actionbook/actionbook --skill deep-research --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From actionbook/actionbook, a repository with 1,584 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.