Awesome AI Sources
PassAudited by ClawScan on May 15, 2026.
Overview
This appears to be a straightforward public RSS feed fetcher, with the main things to notice being outbound requests to Agentic Brew and optional user-approved recurring schedules.
This skill looks reasonable for fetching public AI RSS feeds. Before installing, be aware that it will make outbound requests to www.agenticbrew.ai, may run a local Python fetch/parse command, and can propose recurring schedules if you choose daily or weekly delivery.
Publisher note
1. What it does: fetches curated AI content from Agentic Brew's public RSS feeds and returns a compact list (markdown or JSON) for the calling agent. 2. Network access - Egress hosts: https://www.agenticbrew.ai only (specifically /feed/<feed>.xml — 11 known endpoints). - No authentication. All endpoints are publicly accessible RSS 2.0; no API keys, tokens, or credentials are sent or stored. - No ingress. The skill makes outbound HTTPS only. 3. Data flow - Input: user-supplied args (feed name, limit, query keyword, format). All sanitized into the fetch URL or post-parse filter. - Output: plain text (markdown list) or JSON to stdout. No files written, no caches, no persistent state. 4. Permissions required - Outbound HTTPS to www.agenticbrew.ai. - No filesystem access beyond stdin/stdout. - No shell injection vectors: user-supplied args are passed to Python as sys.argv, not concatenated into shell. 5. Risk summary: Low. Read-only consumer of a public RSS feed. No code execution from feed contents (XML is parsed, not interpreted). No credential handling. No write paths. The worst-case failure mode is a stale or unreachable feed returning no items. 6. Source & maintainer - Source: https://github.com/sunxiayi/awesome-ai-sources/tree/main/plugins/ai-news-radar - Maintainer: sunxiayiatwork@gmail.com - License: MIT
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Using the skill will contact Agentic Brew and display content or links from its public feeds.
The skill performs outbound HTTP requests to a third-party site, but this is disclosed, read-only, and central to the stated RSS-fetching purpose.
Pulls items from the Agentic Brew public RSS endpoints and returns them as a clean list. No auth, no scraping — just an HTTP GET
Use it when you are comfortable fetching public feed data from www.agenticbrew.ai, and treat external links or feed content as informational rather than trusted instructions.
The host agent may execute a local Python command to fetch and parse the RSS feed.
The instruction-only skill appears to rely on a local Python one-liner for fetching and parsing. This is purpose-aligned, but the full command is not visible in the provided truncated excerpt.
Run the fetch + parse one-liner below. It uses the Python stdlib only (`urllib`, `x
If prompted to approve command execution, review that it only fetches the intended Agentic Brew feed and uses the expected Python standard-library parsing.
If you accept a schedule, the agent may keep fetching these feeds on a daily or weekly cadence.
The skill can lead to recurring automated fetches, but the artifact says to propose or offer scheduling rather than silently creating it.
`daily` — fetch now AND propose setting up a recurring task. In Claude Code, suggest the `/schedule` skill (cron) or `/loop` (interval).
Only enable recurring schedules if you actually want periodic fetches, and confirm how to view, modify, or cancel the schedule.