Article Summary Card
v1.0.2Summarize articles or webpages into structured bundles exported as JSON, Markdown, HTML, and PNG with consistent layout and tagging.
⭐ 0· 96·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match what the skill implements: extracting article text, summarizing (prompts provided), and rendering outputs. Declared runtime binaries (python3, curl, Chrome) and Pillow are appropriate for fetching, parsing, and producing screenshots; no unrelated credentials or tools are requested.
Instruction Scope
SKILL.md and code restrict actions to extracting article text (curl or local file), doing a two-pass in-session summary using the provided prompt files, and rendering outputs locally (Markdown/HTML/PNG). The Python scripts call curl and headless Chrome for fetching and screenshotting, which is consistent with the stated workflow. There are no instructions or code paths that attempt to read unrelated config, environment secrets, or post data to hidden external endpoints.
Install Mechanism
No install spec (instruction-only in registry); included install helper is a local copy operation into ~/.openclaw/workspace/skills which is reasonable. No remote downloads or archive extraction from arbitrary URLs are used.
Credentials
The skill requests no environment variables or credentials. The only runtime requirements are local binaries (python3, curl, Chrome/Chromium) and the Pillow package — proportionate to its function.
Persistence & Privilege
always is false and the skill does not request persistent elevated privileges. The install helper only copies the skill folder into the user's OpenClaw workspace when explicitly run; nothing force-enables the skill or modifies other skills' config.
Assessment
This skill appears coherent and implements what it claims: it fetches a user-supplied URL (via curl), parses the HTML locally, invites the agent to run in-session LLM passes (using the prompt files), and renders a card using headless Chrome and Pillow. Things to consider before installing/running: (1) fetched URLs are controlled by the user — avoid pointing it at internal/metadata endpoints you don’t want accessed from the host (it will perform HTTP requests); (2) the scripts call subprocesses (curl, Chrome) — run in a secure/isolated environment if you’re cautious; (3) the install helper will copy the skill into ~/.openclaw by default — check that path if you have a different workspace; (4) if running on a shared or production host, verify the code (it’s short and readable) and ensure Chrome is run in an environment where network access and filesystem writes are acceptable. No credentials are requested and there are no hidden remote endpoints in the code.Like a lobster shell, security has layers — review code before you run it.
imagevk97d3f0b3tty7k98gmdbxg2w8h83evr2latestvk97d3f0b3tty7k98gmdbxg2w8h83evr2markdownvk97d3f0b3tty7k98gmdbxg2w8h83evr2readingvk97d3f0b3tty7k98gmdbxg2w8h83evr2summaryvk97d3f0b3tty7k98gmdbxg2w8h83evr2workflowvk97d3f0b3tty7k98gmdbxg2w8h83evr2
License
MIT-0
Free to use, modify, and redistribute. No attribution required.