Simplified search for news, earthquakes, Iran, tech, finance, layoffs. Bilingual (EN/中文). Flexible time. Data from World Monitor (fetch then query locally)....

This skill appears to do what it claims: it uses a headless Chromium (via Playwright) to load worldmonitor.app, intercept JSON API responses, save them to ~/.openclaw/intel-data/, and produce a local summary that the query script reads. Before installing/running: (1) Review package.json and the two scripts if you want extra assurance (they're included here); (2) Expect npm install and Playwright to download packages and a Chromium binary — run these in an environment where large downloads are acceptable; (3) The captured-data.json and intel-brief.md may contain sensitive or identifiable content (e.g., Telegram messages or links) — restrict the data directory permissions or set INTEL_DATA_DIR to a safe location; (4) The fetch step contacts worldmonitor.app and will store whatever JSON the site returns; if you are concerned about private/session state, run it in a sandboxed account/container — Playwright's newContext is fresh and should not pick up your browser cookies, but running in an isolated environment reduces risk; (5) If you need further assurance, run npm install in a transient VM/container or inspect installed node_modules before executing fetch.

Type: OpenClaw Skill Name: intel-search Version: 1.0.0 The OpenClaw AgentSkills bundle 'intel-search' is designed to fetch data from 'https://worldmonitor.app' using Playwright (scripts/fetch.mjs), store it locally in '~/.openclaw/intel-data/', and then allow the AI agent to query this local data (scripts/query.mjs). All network requests are confined to the stated 'worldmonitor.app' domain, and data is saved locally without any evidence of exfiltration to unauthorized external endpoints. Instructions in SKILL.md and output from scripts/query.mjs (e.g., 'LANG: en') are for guiding the AI agent's presentation and localization, not for malicious prompt injection. While the skill uses powerful tools like Playwright and allows broad 'Bash(node:*)' permissions, these are necessary for its stated purpose and are not leveraged for any intentional harmful behavior such as credential theft, persistence, or unauthorized remote control.