| name | copilot-web-fetch |
| description | Fetch and read the full contents of a specific web page (URL) as clean Markdown using Copilot Plus. Use when the user shares a link or asks you to open, read, or summarize a particular page — not for an open-ended web search. Requires an active Copilot Plus license; without it, use your own fetch tool instead. |
| license | Copilot Plus |
| metadata | {"copilot-enabled-agents":"claude, codex, opencode","copilot-builtin-version":"5"} |
Copilot web fetch
Fetch a web page's contents as Markdown through Copilot Plus.
How to run
Find the absolute path to this SKILL.md file on disk, then run the script next
to it that matches the operating system. No extra runtime is needed — sh
(macOS/Linux) and cmd/PowerShell (Windows) are always present.
On macOS or Linux:
sh "/absolute/path/to/this/skill/directory/web-fetch.sh" "<url-to-fetch>"
On Windows, run the .cmd wrapper. In PowerShell you must prefix it with the
call operator & (PowerShell treats a quoted path on its own as a string and
won't run it); from cmd, run the quoted path without the &:
& "/absolute/path/to/this/skill/directory/web-fetch.cmd" "<url-to-fetch>"
Both print the result to stdout.
If Copilot Plus is not active
If the script exits saying Copilot Plus is unavailable, do NOT retry it. Do what
the message says: fall back to your own equivalent built-in capability to handle
the request when you have one (otherwise tell the user it's unavailable) — never
refuse or block the user. Only mention upgrading or renewing Copilot Plus when
the script's message explicitly invites it, and keep any such note short and
friendly.
