LogoAwesome Skills
  • Search
  • Category
  • Tag
  • Blog
LogoAwesome Skills
LogoAwesome Skills

Discover Open-Source Agent Skills for AI Coding Assistants

Product

  • Search
  • Category
  • Tag
  • Blog

Resources

  • Claude Skill Docs
  • Antigravity Skills Docs

Tools

  • Claude Code
  • OpenCode
  • Cursor
  • Codex
  • Antigravity

Company

  • Privacy Policy
  • Terms of Service
  • Sitemap

©2026 Awesome Skills. All rights reserved.

Privacy PolicyTerms
Back to Skills

kstartup-search

공공데이터포털 창업진흥원 K-Startup Open API(15125364)로 통합 공고 사업 정보·지원사업 공고·창업 콘텐츠·통계보고서를 k-skill-proxy 경유로 조회한다. 검색 전용. 돌쇠에서는 공식 표면을 통한 후속 액션까지 진행한다.

6,898stars810forksUpdated 8/3/2026
Data Science & ML#search#korea#startup#public-api#python

Security Assessment

Safe(90/100)
Security Score90/100

About kstartup-search

K-Startup Search is a Korean-locale business skill that queries the Korean public-data K-Startup Open API (Korea Institute of Startup and Entrepreneurship Development, dataset 15125364) for integrated business announcements, support-program notices, startup content, and statistical reports, routed through a k-skill-proxy. It solves the problem of filtering government startup-support programs (for example finding programs closing this month, region-filtered open recruitments, or the latest statistical reports) from an API whose server-side filtering is inconsistent. It is search-only; any actual application is left to the official K-Startup site.

The SKILL.md is a CLI stub pointing to an npx instruct command, while the bundled scripts/run_kstartup.py (Python standard library only) exposes four subcommands (business-info, announcements, contents, statistics) with rich filters for year, category, region, applicant target, recruitment status, and date ranges. It fetches a small bounded page first to confirm the schema, then applies client-side filtering for fields the upstream does not reliably filter server-side (such as region, applicant target, and startup period), normalizing region names against a 17-province mapping table and attaching a client_filter block to the response. It cites sources via endpoint name, page, and detail URLs. By default it uses a hosted proxy so no user API key is needed; a --direct mode calls the API directly using a KSKILL_KSTARTUP_API_KEY resolved through vault, environment variables, or a 0600 dotenv, and a --dry-run mode prints the request without a key.

It targets founders, startup-support seekers, and researchers in Korea. Credentials are handled through official public-data-portal keys and OS-level vault/dotenv storage with no exfiltration, and the skill enforces hard rules against payments, submissions, and storing plaintext credentials, and documents API error and rate-limit codes.

FAQ

What can I search?

Four endpoints of the K-Startup Open API: integrated business information, support-program announcements (the most useful), startup content like notices and news, and statistical reports.

Do I need an API key?

Not for normal use: by default it goes through the hosted k-skill-proxy, which holds the key. A personal KSKILL_KSTARTUP_API_KEY is only needed for --direct mode, requiring an approved (free, auto-approved) usage application for dataset 15125364 on the public data portal.

How are credentials stored?

For --direct mode, keys are resolved in order from a provisioned vault capability or injected environment variables, host vault, then a 0600-permission ~/.config/k-skill/secrets.env dotenv; general queries only read the proxy URL and never handle the key locally.

Why does it filter results a second time on the client?

The upstream API does not reliably apply some filters (like region) server-side, so the helper re-filters fields such as region, applicant target, and startup period, normalizing region names and adding a client_filter block to the response.

How fresh is the data?

The official service design specifies roughly once-daily updates, so deadlines and recruitment status may lag; the skill advises confirming final status on the announcement's detail page and treats it as search-only, not application automation.

All Files

6 files
tests/__init__.py0.0 KB
View
instruction.md7.6 KB
View
scripts/run_kstartup.py16.1 KB
View
skill.json0.5 KB
View
tests/test_run_kstartup.py14.2 KB
View
SKILL.md1.4 KB
View

Install kstartup-search

Download and extract the skill files to your .claude/skills/ directory.

Quick Setup:

  1. Copy the skill folder to .claude/skills/
  2. Claude will automatically detect and use the skill

Repository

nomadamas/k-skill

Related Skills

tigris-security-access-control

3

rebuttal-writing

362

building-blocks

189

docs-search

189