Karpathy Curated RSS Brief
v1.0.2Fetch articles from Karpathy's curated 93 RSS feeds and generate a Chinese tech daily newsletter. Triggers: RSS 日报、RSS 简报、karpathy-curated-rss-brief、每日简报、Kar...
⭐ 0· 360·3 current·3 all-time
byME柒的旅途@mesevenjourney
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 the included code and instructions: a Python fetch script plus an output template and guidance to generate a newsletter. No credentials, binaries, or unrelated dependencies are requested. Minor note: SKILL.md lists a local OPML path (SKILL_DIR/hn-popular-blogs-2025.opml) but the actual script fetches the OPML from a GitHub Pages URL — functionally equivalent but slightly inconsistent in wording.
Instruction Scope
SKILL.md confines actions to fetching the OPML/feed list, running the included fetch_feeds.py, using a WebFetch tool to retrieve article pages, producing a Markdown file, and validating encoding. It does not instruct reading other system files, environment variables, or sending data to third-party endpoints other than the OPML host and discovered feed URLs. One stylistic oddity: it requires deriving SKILL_DIR from conversation context rather than filesystem search, but this is an operational preference, not a secret-reading directive.
Install Mechanism
No separate install spec; the script relies on PEP 723 inline metadata and the platform's 'uv run --script' to install dependencies (feedparser, aiohttp). This is proportionate to the task (HTTP fetch + feed parsing). As with any runtime install, the platform will download packages from package indexes — expected for this functionality.
Credentials
The skill declares no required environment variables, no credentials, and no config paths. The script and SKILL.md do not attempt to read any other env vars or secrets. Network access is necessary and limited to the OPML URL and discovered feed/article URLs.
Persistence & Privilege
always is false and the skill is user-invocable. It does not request persistent agent-wide privileges or attempt to modify other skills. It writes output to the current working directory as part of its normal behavior (documented).
Assessment
This skill appears to do exactly what it says: fetch a curated OPML, retrieve recent feed entries, read article pages, and produce a Markdown newsletter. Before installing, consider: 1) The OPML is fetched at runtime from mesevenjourney.github.io — review that OPML (or host a local copy) to ensure the feed list is what you expect, since the list determines which external sites the skill will visit. 2) 'uv run --script' will install Python packages (feedparser, aiohttp) at run time — if you restrict package installs, run the script in a controlled environment. 3) The skill writes files to the agent's current working directory; ensure that path is acceptable. 4) The WebFetch step will retrieve full article pages from arbitrary external sites listed in the OPML — be aware of network traffic and privacy implications. If you want stronger guarantees, inspect the OPML and the included fetch_feeds.py (already present) and, if desired, replace the remote OPML with a vetted local copy.Like a lobster shell, security has layers — review code before you run it.
aivk9726rh49wgrpbrxvycppz3w2n82mgjschinesevk9726rh49wgrpbrxvycppz3w2n82mgjslatestvk973xeh13qbxyszfj691rqz4c582n1xbnewslettervk9726rh49wgrpbrxvycppz3w2n82mgjsproductivityvk9726rh49wgrpbrxvycppz3w2n82mgjsrssvk9726rh49wgrpbrxvycppz3w2n82mgjs
License
MIT-0
Free to use, modify, and redistribute. No attribution required.