How Tessl works

Explore how Tessl makes context available to agents, covering the different context types, and how agents receive and use this context.

Tessl is an agent enablement platform that provides a package manager and evaluation tools for managing context. This guide introduces the core types of context you can manage with Tessl, how agents use that context, and the built-in tools available for evaluating it.

Types of context

A skill encapsulates a specific capability or workflow that an agent can execute. A tile is a structured container of context that can include a combination of skills, documents, and rules. The table below explains when to use each type of context and how agents access and apply them.

Type	When Used	Purpose
📚 Docs	On-demand (lazy pull)	Current library knowledge via MCP
📋 Rules	Always (eager push)	Your team's coding standards
🔧 Skills	When relevant (lazy push)	Systematic procedural workflows

The Tessl package manager lets you install and manage both tiles from the Tessl registry and skills from GitHub. See the quick start guide to start using the package manager.

Evaluation tools

Tessl provides built-in tools to evaluate the quality and effectiveness of the context contained in a tile. These tools help you understand how well your context supports agent behavior and where it can be improved. Currently, Tessl supports evaluation for docs and skills. Evaluation tools allow you to generate realistic scenarios to test your context and assess how changes to that context improve agent performance over time. This makes it easier to validate assumptions, compare iterations, and catch issues early. See the common workflows evaluating skills and evaluating documentation for a step-by-step guide on how to use the tools. Tessl has evaluated more than 2,000 public skills against established best practices. Explore the results in the Tessl Registry at tessl.io/registry.