Openclaw Skills for finding, verifying, downloading, and extracting official China CDC health reports with strict source and workflow controls.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install cdc-data
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install cdc-data using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Find China CDC Health Report?
Find China CDC Health Report is an Openclaw Skills workflow for locating specific official China CDC publications, validating their authenticity, and performing only the requested action such as locating, downloading, or extracting content.
It is designed for precise report discovery across influenza weekly reports, respiratory sentinel reports, COVID-19 monthly updates, and infectious-disease risk assessments, with strict rules for source trust, carrier verification, and task-local execution.
Find China CDC Health Report Use Cases
- Find the latest China CDC influenza weekly report or a specific week/issue.
- Verify whether a report PDF or detail page is officially hosted by China CDC.
- Download a verified official report PDF to a local artifact directory.
- Extract report text, tables, charts, and metadata from HTML or PDF sources.
- Compare two official reports when methodology and period compatibility are confirmed.
- Resolve ambiguous report requests by returning the matching official report type options.
- Use Openclaw Skills for source-verified public health reporting without relying on third-party mirrors or search engines.
How Find China CDC Health Report Works
- Parse the user request into normalized fields such as report type, year, week, month, issue number, publish date, reference date, latest, action, and output format.
- Select exactly one supported adapter from the China CDC source registry based on the requested report family.
- Load only the reference files needed for the requested action, such as matching rules, report schema, PDF extraction rules, or comparison metadata.
- Locate the official report by opening the adapter index URL, verifying host and path scope, and following semantic pagination while tracking visited canonical URLs.
- Verify the report carrier and identity using the correct method for the source type: detail page plus PDF for HTML-with-PDF, detail page only for HTML, or direct PDF for PDF-native reports.
- If the user requested download, transfer only verified bytes with the official downloader and save them atomically to the artifact directory.
- If the user requested extraction, generate exactly one extracted.json artifact and populate it with verified content from HTML or PDF sources.
- Use vision only when explicitly requested and only for pages already marked in the extraction artifact; then return a concise status with verified URLs, period data, warnings, and saved paths.
Find China CDC Health Report Setup
- Ensure the Openclaw Skills runtime can access the browser and filesystem tools required by the skill.
- Verify that the repository includes the referenced scripts and reference files under the expected paths.
- Use the unified runner for locate, download, or extract actions.
python3 scripts/fetch_report.py \
--report-type <id> (--latest | --year <year> (--week <week> | --month <month>)) \
[--issue <issue>] --action <locate|download|extract>
- For explicit download requests, use the official downloader with a verified PDF URL and referer.
python3 scripts/download_official_document.py \
--url <verified-pdf-url> \
--referer <verified-detail-or-index-url> \
--allowed-path-prefix <adapter-prefix> \
--report-type <id> --year <year> (--week <week> | --month <month>) \
[--issue <issue>] --output-dir <artifact-directory>
- Keep execution isolated per task, and do not persist browser memory, cookies, or execution traces.
- Confirm that only allowed China CDC URLs are used and that no search engine, mirror, or CAPTCHA bypass is involved.
Find China CDC Health Report Data Schema & Taxonomy
This skill organizes data around a strict report identity and artifact model used by Openclaw Skills.
| Field | Purpose | Notes |
|---|---|---|
| report_type | Selects the report family | influenza, respiratory, covid, global risk, or China risk |
| year | Reporting year | Must be validated against the official title |
| week | Weekly period index | Valid only for weekly reports; influenza also supports issue lookup |
| month | Monthly period index | Used for monthly reports |
| issue_number | Official issue identifier | Must match exactly when used |
| publish_date | Publication date on the official source | Used for verification, not as the primary period matcher |
| reference_date | Candidate period mapping | Requires title verification |
| latest | Flag for latest-record lookup | Uses list-page inspection rules |
| action | Requested operation | locate, download, or extract |
| output_format | Output control | Used when the workflow supports structured output |
Generated artifacts and outputs:
artifacts/<report-id>/extracted.json: the only parsing artifact allowed for extraction.- Downloaded PDFs: saved only when explicitly requested by the user and managed by the downloader.
- Temporary extraction PDFs: allowed only in a task-temporary directory and must be deleted after use.
Extraction content taxonomy:
content: verified main text, tables, figure titles, and image URLs for HTML sources.- Per-page PDF data: native text, tables, and page order for PDF sources.
requires_vision: marks pages that need visual interpretation.vision_pages: lists only the page numbers requiring vision.approximate=true: marks chart values estimated through vision.
Validation and matching rules:
- Validate week ranges 1-53, month ranges 1-12, and positive issue numbers.
- Match year/week or year/month from the official title, not from the URL.
- Return
not_foundonly after a complete and healthy search. - Keep all browser-session state task-local and non-persistent.
Find China CDC Health Report Advanced Features
- Supports one-process locate/download/extract execution with a unified runner for fewer browser round trips.
- Uses strict source registry selection to ensure only one adapter and one official report family are loaded.
- Enforces carrier-aware handling for HTML, HTML-with-PDF, and direct-PDF sources.
- Includes fallback-to-HTTP read-only verification when browser navigation or verification fails.
- Supports reusable extraction artifacts so valid results can be retained without unnecessary reprocessing.
- Enables optional vision-based interpretation only for explicitly marked PDF pages that cannot be read natively.
- Provides comparison support for compatible reports when methodology, geography, period, and denominators align.
- Uses task-local sessions and artifact hygiene to prevent persistence of cookies, browser memory, or intermediate render data.
- Helps Openclaw Skills users safely work with authoritative public health sources while avoiding third-party or unregistered content.
SKILL.md
Loading
Related Openclaw Skills
Luoshenfu Literary Research
Openclaw Skills for researching and writing a scholarly Simplified Chinese analysis of Ode to the Goddess of Luo River, with citations and annotations.
Zhiyanzhai Research Article Generator
Openclaw Skills for generating a long-form, citation-backed Simplified Chinese research article on the identity of Zhiyanzhai in Dream of the Red Chamber.
find-cargo-for-vessel
Openclaw Skills for vessel cargo matching finds domestic or international cargo boards by current port, destination port, and ship capacity.
招采萝卜坑识别专家 · 招标文件量身定制扫描
Openclaw Skills for scanning full tender documents to identify hidden "carrot pit" clauses, exclusion signals, and compliance risks.
Image to 3D Part STL Builder
Openclaw Skills converts dimensioned images into confirmation-ready SVG blueprints and accurate STL models for fabrication.
READ() - Reading to Action Transformation
READ() transforms reading materials into structured insight and action, turning content into reasons, questions, evidence, next steps, and review points.
Featured*
