hkust-gz-faculty-intel
Crawl and integrate HKUST(GZ) public faculty profile data by Hub and one or more Thrust areas. Use when Codex needs to collect faculty lists, emails, titles, profile URLs, identifiers, jobs, acting he
SKILL.mdskills/hkust-gz-faculty-intel/SKILL.md
GitHubHKUST-GZ Faculty Intel
Overview
Use this skill to crawl public HKUST(GZ) Faculty Profile data at Hub and Thrust granularity. The default path uses public JSON endpoints behind facultyprofiles.hkust-gz.edu.cn, not login cookies, browser automation, or private APIs.
Quick Start
List supported Hubs and Thrusts:
python skills\hkust-gz-faculty-intel\scripts\hkust_gz_faculty.py list-thrusts
Crawl one Hub plus one Thrust:
python skills\hkust-gz-faculty-intel\scripts\hkust_gz_faculty.py crawl `
--hub "Information Hub" `
--thrust "Artificial Intelligence" `
--output-dir output\hkust_gz\info_ai
Crawl several Thrusts under one Hub:
python skills\hkust-gz-faculty-intel\scripts\hkust_gz_faculty.py crawl `
--hub INFOHUB `
--thrust "Artificial Intelligence" `
--thrust "Data Science and Analytics" `
--output-dir output\hkust_gz\info_ai_dsa
Crawl every Thrust under a Hub:
python skills\hkust-gz-faculty-intel\scripts\hkust_gz_faculty.py crawl `
--hub "Systems Hub" `
--all-thrusts `
--output-dir output\hkust_gz\systems
Use official Thrust URL or code directly:
python skills\hkust-gz-faculty-intel\scripts\hkust_gz_faculty.py crawl `
--thrust-url "https://facultyprofiles.hkust-gz.edu.cn/thrust-faculties?code=10011A10000000000H28" `
--output-dir output\hkust_gz\ai_by_url
Preview requests without network crawling:
python skills\hkust-gz-faculty-intel\scripts\hkust_gz_faculty.py crawl `
--hub INFOHUB `
--thrust "Artificial Intelligence,Data Science and Analytics" `
--dry-run
Workflow
- Resolve the user's target into official Hub and Thrust codes. Use
list-thrustswhen the request says "港科广某个 hub/系/学域" but names are abbreviated. - Prefer
crawl --hub ... --thrust ...for exact Hub/Thrust subsets. Use repeated--thrustor comma-separated names for multiple Thrusts. - Use
--include-detailsonly when the user needs biography, projects, publications, resumes, or research-interest details. It calls one profile-detail endpoint per person and can produce large JSON. - Keep raw API captures separate from normalized exports. Do not commit crawl outputs unless the user explicitly asks.
- If the public endpoint returns empty data for a known Thrust, retry with
--source page; report endpoint drift instead of inventing data.
Output
Each crawl run writes:
hkust_gz_faculty_manifest.json: selected Hub/Thrusts, source URLs, counts, and run metadata.hkust_gz_faculty_profiles.json: normalized person-Thrust rows with selected jobs and source provenance.hkust_gz_faculty_profiles.csv: flat table for spreadsheet review.raw/*.json: raw public endpoint captures for audit and reruns.
Read references/output-schema.md before merging this output into another dataset.
Scope
This skill is specialized for HKUST(GZ):
- Main Hubs: Function, Information, Systems, Society.
- Additional public units present in the same profile site may be listed, but Hub/Thrust crawling should default to the four main Hubs unless the user asks otherwise.
- It does not crawl HKU, CUHK, CityU, PolyU, or HKUST Clear Water Bay faculty directories.
- It does not bypass login, CAPTCHA, rate limits, or private data controls.
Resources
scripts/hkust_gz_faculty.py: standard-library crawler and exporter.references/hub-thrust-map.md: built-in Hub/Thrust code map and aliases.references/output-schema.md: normalized fields and merge policy.