Index of every page on this site: llms.txt
All skills
vercel avatar

/try-it

@86694bd
by vercelvercel/board147 stars
13
Run it yourself

Find a way to try your changes to a repo for real, by using what it builds, and keep what you learn for the next thread. Says what can and can't reach your sandbox, so you ask a person only for what can arrive. Use when you start on a repo you haven't run, and when a change needs more than its tests.

Use this Skill: https://skilld.dev/gh/vercel/board/try-it

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈77 tokens for metadata: the name and description. ≈416 when used: this file.

Description uses 3.9% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈310 characters, or 3.9%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Try your work for real

Tests tell you some of it. Using the thing tells you the rest: open the page, run the command, call the endpoint. For each repo you work on, look for a way to do that yourself.

How is yours to work out from what's there. Every repo is different, and some have no way you can reach. That's fine: say what you couldn't try, and carry on.

What the repo won't tell you:

  • What's in your sandbox is the repo, what it installs, and what you make there: a local database, seed data, an account you sign up for on a server you started.
  • A service's key reaches your requests, never your sandbox (the add-service skill). Nothing else comes in. A variable in this factory's Vercel project isn't in your environment, so don't ask for a password, a client secret, or a connection string to be put there: it would never arrive.
  • A person can do what happens outside your sandbox: add a service's key, make an account or a test project, change a setting, try one step and say what they saw. Get as far as you can first, then ask once, for all of it.
  • A protected preview is closed to you. It asks for a Vercel login, which your browser can't do.
  • In a thread you start yourself, your GitHub access covers every repo the board brings in from one org, this factory's own. A repo in another org can't be cloned there: work on it from the thread of one of its pull requests or issues.

When you find a way, show it once in the thread (a screenshot, what the command printed), and write it down in the repo's AGENTS.md: what to run, what it needs, and what can't be tried. The next thread starts from there.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No third-party reports yet.

Signed by skilld at 86694bd. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 12 hours ago.

Activeupdated 3 days ago

README badge

README badge for vercel/board/try-it