OpenClaw News

Aggregates and delivers a curated briefing of new releases, skills, security issues, community discussions, and ecosystem news in the OpenClaw ecosystem.

This skill is functionally coherent as a news-aggregation tool, but the registry metadata fails to list runtime dependencies and credential requirements. Before installing or running it: (1) Verify you are comfortable granting the skill access to your local 'gh' CLI (it will use whatever GitHub auth gh has configured). If you don't want it to use your personal GH token, don't run the scripts or run them in an environment with limited gh credentials. (2) Confirm whether your agent's web_search/Brave Search integration requires an API key and whether you're willing to let the agent run those network searches. (3) Ensure you trust the clawdhub CLI output or run the skill in a sandbox first; clawdhub may query registries and could surface unexpected data. (4) Review the state/ files (pending_searches.json, raw_data.json) to see what queries and results are produced; they are stored locally and could contain noisy or sensitive registry entries. (5) Because metadata omitted required tools/creds, treat this as a transparency issue — if you need full assurance, ask the publisher to update metadata to declare gh/clawdhub/Brave Search requirements and any needed env vars, or run the scripts manually in a controlled environment.

Type: OpenClaw Skill Name: openclaw-news Version: 1.0.0 The OpenClaw Ecosystem News skill is designed to aggregate news from specified sources like GitHub, ClawdHub, and web searches. All scripts (`collect_news.sh`, `format_briefing.sh`, `agent_collect.py`) and the `SKILL.md` instructions are transparent, directly supporting the stated purpose. The skill uses standard CLI tools (`gh`, `clawdhub`) and Python for data collection and formatting, with file system operations limited to its own state directory. There is no evidence of prompt injection, credential theft, data exfiltration, persistence mechanisms, or execution of arbitrary untrusted code.