Guide

How to use @clarvis/agent-tools in practice: wire it into an agent loop, drive the core directly for a custom transport, run it read-only, and tune its output and input limits. Every page here is task-oriented; the Reference has the exhaustive contracts.

Pick your level

The factory — for embedding in an agent loop

createAgentTools(options) resolves a config once and returns listTools() / callTool(). This is what most consumers want: advertise the surface to a model, then dispatch the tool calls the model makes.

→ Embed it in an agent loop

The core — for building your own transport

The building blocks — resolveConfig, dispatch, listTools, and the raw tool registry — are exported so you can expose the tools however you like: an HTTP endpoint, a CLI, or a test harness.

→ The core API

Then tune it

Read-only mode — expose only read_file / list_dir / glob / grep, and drop the mutating tools entirely.
Limits & spill — how output is bounded, how oversized input is refused, and how bash overflow spills to disk.

Guide ​

Pick your level ​

The factory — for embedding in an agent loop ​

The core — for building your own transport ​

Then tune it ​

See also ​

Guide

Pick your level

The factory — for embedding in an agent loop

The core — for building your own transport

Then tune it

See also