Collect, filter, classify AI industry news, generate Chinese titles and summaries, and export Excel and Word briefs based on company lists and sources.

Things to check before running: - Provide or inspect config.yaml: The code expects a config.yaml (Settings raises if missing). Review config.default.yaml included in the repo and either provide a config.yaml or run with the correct config path. - Review source_config.xlsx / web_source_config.xlsx: The skill will fetch all URLs listed there. Do not include sensitive intranet/internal URLs unless you want them crawled. - External summarization: If you set ARK_API_KEY, the skill will send article content to Volcengine Ark (Doubao) for generation of Chinese titles/summaries. Do not supply that key if scraped content is sensitive or you do not consent to external processing. - Local data written: The skill will create output directories, logs, and an SQLite database (paths controlled via config). Run initially in an isolated environment or container if you want to inspect behavior first. - Dependencies: Install Python packages from requirements.txt in a virtualenv to avoid affecting system packages. - Code review: If you need higher assurance, open the remaining service files (ai_summary_service.py, briefing_service.py, etc.) to confirm the exact network endpoints and what fields are sent to external APIs. Given the above mismatches (missing config mention and the potential for external data transmission), treat this skill as suspicious until you confirm configuration and data flows are acceptable.

Type: OpenClaw Skill Name: ai-news-workflow Version: 0.1.3 The 'ai-news-workflow' skill bundle is a well-structured news aggregation and processing tool. It collects data from RSS feeds and web pages (app/crawlers/), processes it through various services for classification, deduplication, and company matching, and utilizes the Volcengine Ark (Doubao) LLM API for generating summaries and identifying key events (app/services/ai_summary_service.py). The code uses standard libraries like pandas, openpyxl, and requests, follows secure database practices with parameterized SQLite queries (app/repositories/news_repository.py), and handles API keys via environment variables. No evidence of malicious intent, data exfiltration, or prompt injection attacks was found.