Naver news Search

Search Korean news articles using Naver Search API. Use when searching for Korean news, getting latest news updates, finding news about specific topics, or preparing daily news summaries. Supports relevance and date-based sorting.

Security Scan

The skill's code, instructions, and required environment variables are consistent with a Naver News search utility and request only the credentials needed to call the official Naver Search API.

Purpose & Capability

The skill name/description, included Python script, and required env vars (NAVER_CLIENT_ID, NAVER_CLIENT_SECRET) align with calling Naver's openapi.naver.com news search endpoint. Requiring python3 and the two Naver credentials is proportionate to the stated purpose.

Instruction Scope

The SKILL.md and examples provide CLI usage and automation guidance. They suggest storing runner state in memory/news-state.json for cron automation (a convenience for automation workflows). The runtime instructions and script only read env vars and call the official Naver API; they do not instruct reading unrelated system secrets. If you enable the automation example, the agent will need file access to whatever state file you choose.

Install Mechanism

No install spec or external downloads are used; this is instruction-only with an included Python script. Risk is low — the script uses standard library modules and makes network calls to Naver's official API.

Credentials

Only NAVER_CLIENT_ID and NAVER_CLIENT_SECRET are required, which are the expected credentials for Naver's API. The code uses those credentials only to set HTTP headers to openapi.naver.com. No unrelated credentials or secrets are requested.

Persistence & Privilege

always:false and no unusual privilege escalation are requested. The SKILL.md recommends adding credentials to OpenClaw config locations (sandbox or host), which is normal; ensure you store credentials in the appropriate, least-privileged config (prefer sandbox) and do not expose them to other agents.

Assessment

This skill appears coherent: it needs only python3 and your Naver app credentials to call Naver's official news search API. Before installing: (1) create a Naver developer application and use its Client ID/Secret; (2) store those credentials securely (prefer sandbox-scoped env over global host env) and limit which agents receive them; (3) be aware the script makes network requests to https://openapi.naver.com and will count toward the app's 25,000/day quota; (4) if you enable automated runs described in examples (cron/state files), ensure the agent's workspace and any state files (e.g., memory/news-state.json) do not contain other secrets; (5) if you have strict compliance requirements, review the included Python script yourself — it is short and uses only standard libraries.

Search Korean news articles using the Naver Search API.

Quick Start

Use the provided script to search news:

python scripts/search_news.py "검색어" --display 10 --sort date

Options:

• --display N: Number of results per page (1-100, default: 10)
• --start N: Start position for pagination (1-1000, default: 1)
• --sort sim|date: Sort by relevance (sim) or date (date, default: date)
• --after DATETIME: Only show news published after this time (ISO 8601 format, e.g., 2026-01-29T09:00:00+09:00)
• --min-results N: Minimum number of results to fetch (enables auto-pagination)
• --max-pages N: Maximum number of pages to try when auto-paginating (default: 5)
• --json: Output raw JSON instead of formatted text

Setup

Environment Variables

Required credentials from https://developers.naver.com/:

NAVER_CLIENT_ID=your_client_id
NAVER_CLIENT_SECRET=your_client_secret

Configuration locations:

• Sandbox (default): Add to agents.defaults.sandbox.docker.env in OpenClaw config
• Host: Add to env.vars in OpenClaw config

Getting API Credentials

1. Visit https://developers.naver.com/
2. Register an application
3. Enable "검색" (Search) API
4. Copy Client ID and Client Secret
5. Add credentials to appropriate config section (see above)

Common Use Cases

Latest news on a topic

python scripts/search_news.py "AI 인공지능" --display 20 --sort date

Search with relevance ranking

python scripts/search_news.py "삼성전자" --sort sim

Filter by time (only recent news)

# News published after 9 AM today
python scripts/search_news.py "경제" --display 50 --sort sim --after "2026-01-29T09:00:00+09:00"

# News from the last hour (programmatic use)
python scripts/search_news.py "속보" --after "$(date -u -d '1 hour ago' '+%Y-%m-%dT%H:%M:%S%z')"

Auto-pagination for guaranteed minimum results

# Fetch at least 30 results (automatically requests multiple pages if needed)
python scripts/search_news.py "AI" --sort sim --after "2026-01-29T09:00:00+09:00" --min-results 30 --display 50

# Limit to 3 pages maximum
python scripts/search_news.py "게임" --min-results 50 --max-pages 3

How auto-pagination works:

1. Fetches first page (e.g., 50 results)
2. Applies date filter (e.g., 10 results remain)
3. If below --min-results, automatically fetches next page
4. Stops when minimum is reached or --max-pages limit hit

Pagination for more results

# First 10 results
python scripts/search_news.py "경제" --display 10 --start 1

# Next 10 results
python scripts/search_news.py "경제" --display 10 --start 11

Using in Python Code

Import and use the search function directly:

from scripts.search_news import search_news

result = search_news(
    query="경제 뉴스",
    display=10,
    sort="date"
)

for item in result["items"]:
    print(item["title"])
    print(item["description"])
    print(item["link"])

API Details

For complete API reference including response structure, error codes, and rate limits, see:

references/api.md

Notes

• Search queries must be UTF-8 encoded
• Results include <b> tags around search term matches (strip them for clean text)
• Daily limit: 25,000 API calls per application
• link field may point to Naver News or original source depending on availability