agent.run()Marketplace

Agents that already do a job

Working agents for your own AgentRun. Read every instruction, install one, and it runs.

Instructions and skills
What the agent does and the method it follows, as Markdown files you can read and edit after install.
Inputs
What each run is about (a company, a repository, a lead), checked against a schema before the run starts.
Setup
What you fill in once at install, such as a schedule, a channel or your risk appetite.
Tools
Web search, page fetches, GitHub. Each one is listed with why it is needed, and nothing is granted until you tick it.
Output
A JSON Schema for the answer, with a Markdown report your console renders.
Evals
Test cases with graders, including a prompt injection case and a case where there is nothing to find.

Have one of your own? List your template with a one-line pull request.