Agent skills

Introduction

An agent skill is a folder of instructions (SKILL.md and its references) that a coding agent loads when a task calls for it. The reversecontact-cli skill teaches an agent to drive the Reverse Contact CLI: enrich people and companies, fetch by LinkedIn URL or id, search, check freshness for free, find contact emails and phone numbers, follow live jobs and inspect credits.

It is published in the public repository visum-run/skills and works with any agent that reads Agent Skills, such as Claude Code. Each CLI release publishes its skill there.

What the agent learns

• The cheapest command for the input it has. With a profile URL or a Reverse Contact id, it fetches (1 credit) rather than enriches (2 credits), and runs the free rc check first when freshness matters. It enriches only from an email, a name or a domain.
• To check the budget before a batch. It reads the credit balance and the rate limits with rc usage, which is free, before a long run.
• To read results like a program. It relies on the JSON output and the exit codes of rc: a not-found result exits 0 and costs nothing, 3 means no valid API key, 4 a network problem.
• Not to block on live jobs. It starts a live fetch, an email lookup or a phone lookup, keeps the job id, and reads the result later with rc task, instead of waiting in a loop.
• Not to invent search filters. It lists the filters the API accepts with rc search fields people and rc search fields companies, which are free.

Install

The skill needs the rc binary and an API key: see Install and Sign in on the CLI page. An agent is best given the key through the RC_API_KEY environment variable, which writes nothing to disk.

Then add the skill to your agent with the skills CLI:

npx skills add visum-run/skills

To install it by hand, copy the reversecontact-cli folder of the repository into the skills folder of your agent. For Claude Code, that is .claude/skills/ in a project, or ~/.claude/skills/ for every project.

A worked session

You ask the agent: "Who is jane.doe@acme.com, and what is her current role?" With the skill loaded, the agent runs:

# 1. Is a valid API key available? Exit code 3 means no.
rc auth status --json

# 2. An email, no profile URL: enrich, keeping only the fields needed.
rc enrich person --email jane.doe@acme.com \
  --fields firstName,lastName,currentPositionTitle,currentCompanyName,linkedinUrl

With --fields, rc prints only those fields, flat, without the response envelope, and the agent answers from them. Had no one matched, rc would have exited 0 and charged nothing, and the agent would say so rather than guess.

Skill or MCP server

Both reach the same API, with the same credits.

FAQ

How do I install the Reverse Contact agent skill?

Install the rc CLI and give it an API key, then run npx skills add visum-run/skills. To install it by hand, copy the reversecontact-cli folder of the visum-run/skills repository into your agent's skills folder, such as .claude/skills/ for Claude Code.

Does the Reverse Contact agent skill cost credits?

The skill itself is free. The commands it runs are billed like the Reverse Contact API endpoints they call, and the skill steers the agent to the cheapest one: a free rc check before a 1-credit rc fetch, and rc fetch rather than a 2-credit rc enrich when a profile URL is known.

Should I use the Reverse Contact agent skill or the MCP server?

Use the MCP server for a conversational agent working in a chat, and the skill with the rc CLI for a coding agent that writes or runs scripts. Both call the same Reverse Contact API with the same credits.