by sv-number
Give your AI agent a phone number: order a private number in any country over the API, read the SMS verification code, hand the number back.
Unlocks once the catalog security scan passes (runs nightly).
The deep catalog scan for this skill is still queued. Run an instant dependency check now instead.
# Add to your Claude Code skills
git clone https://github.com/sv-number/skillsGuides for using ai agents skills like skills.
Your agent hits a signup wall: "we sent a code to your phone". It has no phone. This skill gives it one on demand, in the country the service expects, and reads the code back over the API.
One markdown file. No SDK, no dashboard, no human at the keyboard.
> Read the sv-number skill and register me a Telegram account on an Indonesian number.
ordered +62 838 1234 5678
code 123456
account created
Most verification checks look at the country the number came from. An Indonesian marketplace refuses a US number before the SMS is even sent. The agent names the country in the request and that whole class of failure disappears.
The API returns the verification code itself, already parsed out of the message. There is no regex to write per service and nothing to fix when a service rewrites its wording.
Money is held when the number is ordered and charged when the code lands. Nothing in 20 minutes and it returns to the balance by itself.
One activation, one holder. Nobody else receives its codes, which is exactly the part that free shared pools get wrong.
The catalogue call returns, per service and country, the share of codes that arrived for other users and how many numbers are live right now. Where the share is measured, an agent that reads it stops paying for countries that never deliver; where it is not, the size of the live pool tells the same story.
When the account asks for an authenticator, the agent stores the TOTP secret next to its API key and computes the six digits locally, by RFC 6238. The secret never leaves the machine.
curl -O https://raw.githubusercontent.com/sv-number/skills/main/sv-number/SKILL.md
export SVN_API_KEY=your_key_here
The layout follows the Agent Skills specification:
a directory named after the skill, holding SKILL.md and scripts/. Point any runner at
it and it works as is, in Claude Code, Codex or anything else that reads a skill. The key comes from
your profile after
signup. The balance has to be funded:
there is no free tier on this API.
getNumber -> ACCESS_NUMBER:id:phone order a number, money is held
getStatus -> STATUS_WAIT_CODE poll every 3 to 5 seconds
-> STATUS_OK:123456 the code
setStatus 6 -> ACCESS_ACTIVATION done, the charge stays
setStatus 8 -> ACCESS_CANCEL nothing came, money back
Runnable versions: sv-number/scripts/order_number.py
and sv-number/scripts/order_number.mjs.
Receiving codes is the whole job. These numbers do not send SMS and do not take calls, and they are not for banking, payment or government accounts. If your agent needs a permanent number that holds a conversation with people, you want a carrier product instead.
The API key stays in the agent's environment and goes nowhere else. An activation is used
for the service it was ordered under, because that is what keeps delivery rates high for
everyone. Every activation gets closed: setStatus 6 when the code was used, setStatus 8
when nothing arrived. A forgotten activation holds money until it expires.
sv-number/SKILL.mdMIT for the skill and the examples. The service behind it is commercial.
skills is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by sv-number. Give your AI agent a phone number: order a private number in any country over the API, read the SMS verification code, hand the number back. It has 63 GitHub stars.
skills's catalog security scan is still queued. You can run an instant dependency and prompt-injection check now with the "Scan for vulnerabilities" button above.
Clone the repository with "git clone https://github.com/sv-number/skills" and add it to your Claude Code skills directory (see the Installation section above).
skills is primarily written in Python. It is open-source under sv-number on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh skills against similar tools.
No comments yet. Be the first to share your thoughts!