← 返回 Skills 市场
767
总下载
0
收藏
1
当前安装
9
版本数
在 OpenClaw 中安装
/install korean-gov-programs
功能描述
Collect Korean government support programs (TIPS, Small Business, R&D grants) into structured JSONL files. Supports incremental collection with checkpoints.
使用说明 (SKILL.md)
korean-gov-programs
한국 정부지원사업(TIPS, 소상공인, R&D)을 구조화된 JSONL 파일로 수집하는 스킬. 체크포인트 기반 증분 수집으로 중복 없이 안전하게 운영.
수집 소스
| 소스 | 카테고리 | 방식 | 상태 |
|---|---|---|---|
| 기업마당(BizInfo) | 소상공인 | 정적 HTML | ✅ 동작 |
| NIA 한국지능정보사회진흥원 | 정보화사업 | onclick 패턴 | ✅ 동작 |
| 기업마당 기술창업 필터 | 기술창업/R&D | 정적 HTML | ✅ 동작 |
| 소상공인시장진흥공단(SEMAS) | 소상공인 | JS 렌더링 필요 | ⚠️ 스킵 |
| 중소벤처기업부(MSS) | 정부지원 | JS 렌더링 필요 | ⚠️ 스킵 |
| K-Startup | 창업지원 | JS 렌더링 필요 | ⚠️ 스킵 |
| 연구개발특구진흥재단(Innopolis) | R&D | JS 렌더링 필요 | ⚠️ 스킵 |
| 창업진흥원(KISED) | 창업 | eGovFrame 오류 | ⚠️ 스킵 |
JS 렌더링 필요 사이트는 Selenium/Playwright 환경에서 별도 수집 필요.
사용법
# 기본 수집 (./data 디렉토리에 저장)
python3 scripts/collect.py --output ./data
# 커스텀 출력 디렉토리
python3 scripts/collect.py --output /path/to/output
# 수집 현황 확인
bash scripts/stats.sh ./data
JSONL 스키마
{
"title": "사업명",
"category": "소상공인 | 기술창업 | 정보화사업 | R&D",
"source": "수집 출처 기관명",
"url": "상세 페이지 URL",
"amount": "지원 금액 (있는 경우)",
"deadline": "마감일 (예: ~2026-03-31)",
"description": "부가 설명",
"collected_at": "2026-02-19T08:53:00.000000"
}
체크포인트 & 안전 수집
- APPEND 전용: 기존 파일 덮어쓰기 절대 없음
- 중복 방지: title 기준 중복 자동 스킵
- 체크포인트:
.checkpoint.json에 진행 상태 저장 → 재실행 시 이어서 수집 - 딜레이: 요청 간 0.8초 대기 (서버 부하 방지)
출력 파일
data/
├── soho_programs.jsonl # 소상공인 지원사업
├── gov_programs.jsonl # 정부 R&D / 기술창업 지원사업
└── .checkpoint.json # 체크포인트 (자동 생성)
파일 구조
korean-gov-programs/
├── SKILL.md # 이 파일
└── scripts/
├── collect.py # 통합 수집 스크립트
└── stats.sh # 수집 현황 출력
安全使用建议
This appears to be a straightforward scraper: it will perform HTTP requests to government sites and create files in whatever output directory you specify (including a .checkpoint.json). Before running, consider: (1) run it in an isolated directory you control to avoid accidental file placement; (2) respect target sites' robots.txt and rate limits (the script already sleeps 0.8s between requests), and be aware some sources require Selenium/Playwright (skipped by this script); (3) the code optionally reads GOV_SCRAPER_UA to set a User-Agent — harmless but undocumented; (4) the provided collect.py in the review was partially truncated, so if you want higher assurance, open and inspect the entire collect.py for any additional network endpoints or behaviors before executing. If comfortable, run it with a local Python environment only (no elevated privileges) and point --output to a directory you control.
功能分析
Type: OpenClaw Skill
Name: korean-gov-programs
Version: 1.0.8
The skill bundle is a web scraper designed to collect Korean government support program data into JSONL files. The `SKILL.md` provides clear, benign instructions for the AI agent and human users, with no evidence of prompt injection attempts. The `scripts/collect.py` and `scripts/stats.sh` files perform web scraping using `urllib.request` and local file operations (reading/writing JSONL and checkpoint files). All network requests are directed to legitimate Korean government domains, and there is no evidence of data exfiltration to unauthorized endpoints, arbitrary command execution, or other malicious behaviors. The use of environment variables is limited to `GOV_SCRAPER_UA` for user-agent customization, which is benign.
能力评估
Purpose & Capability
The name/description describe collecting Korean government program listings; the repo contains scraping code (scripts/collect.py) and a stats helper (scripts/stats.sh) that align with that purpose. There are no unrelated credentials, binaries, or external packages requested.
Instruction Scope
Runtime instructions and code perform HTTP GETs against government sites, parse HTML, and append structured JSONL records to an output directory, storing progress in .checkpoint.json. The code only reads/writes files under the user-specified output directory and does not access unrelated system files or secrets in the visible portion.
Install Mechanism
No install spec; this is instruction+script only and uses only Python standard libraries and a bash helper. Nothing is downloaded from arbitrary URLs or installed automatically.
Credentials
No required environment variables or credentials are declared. The code optionally reads GOV_SCRAPER_UA to override the User-Agent, which is benign but not documented in SKILL.md (optional only). No sensitive credentials are requested.
Persistence & Privilege
Skill does not request always:true and does not modify other skills or system-wide configs. It creates/updates only its own output files and .checkpoint.json in the user-specified directory (append-only behavior is implemented).
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install korean-gov-programs - 安装完成后，直接呼叫该 Skill 的名称或使用
/korean-gov-programs触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.8
Auto-update 2026-02-25: maintenance & improvements
v1.0.7
Auto-update 2026-02-22: sync with package.json
v1.0.6
Security: add .npmignore, SKILL.md only distribution
v1.0.5
Security: clean publish, no credentials
v1.0.4
이름 수정
v1.0.3
이름 수정
v1.0.2
버전 동기화
v1.0.1
Fix: User-Agent overridable via GOV_SCRAPER_UA env var
v1.0.0
Initial release: TIPS/소상공인/R&D 정부지원사업 JSONL 수집기
元数据
常见问题
Korean Gov Programs 是什么？
Collect Korean government support programs (TIPS, Small Business, R&D grants) into structured JSONL files. Supports incremental collection with checkpoints. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 767 次。
如何安装 Korean Gov Programs？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install korean-gov-programs」即可一键安装，无需额外配置。
Korean Gov Programs 是免费的吗？
是的，Korean Gov Programs 完全免费（开源免费），可自由下载、安装和使用。
Korean Gov Programs 支持哪些平台？
Korean Gov Programs 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Korean Gov Programs？
由 Tomas（@lifeissea）开发并维护，当前版本 v1.0.8。
推荐 Skills