CtrlK
BlogDocsLog inGet started
Tessl Logo

agent-tool-builder

Tools are how AI agents interact with the world. A well-designed tool is the difference between an agent that works and one that hallucinates, fails silently, or costs 10x more tokens than necessary. This skill covers tool design from schema to error handling. JSON Schema best practices, description writing that actually helps the LLM, validation, and the emerging MCP standard that's becoming the lingua franca for AI tools. Key insight: Tool descriptions are more important than tool implementa

Install with Tessl CLI

npx tessl i github:duclm1x1/Dive-Ai --skill agent-tool-builder
What are skills?

38

Does it follow best practices?

Validation for skill structure

SKILL.md
Review
Evals

Evaluation results

99%

-1%

Customer Support Ticket Tool Definitions

Tool schema design with clear descriptions and examples

Criteria
Without context
With context

Descriptions avoid vagueness

100%

100%

Descriptions explain when to use

100%

100%

Parameter descriptions present

100%

100%

Parameter descriptions are specific

100%

100%

Enum constraints used

100%

100%

Input examples included

100%

100%

Required fields declared

100%

100%

Schema types accurate

100%

87%

Minimal toolset

100%

100%

No overlapping tool responsibilities

100%

100%

Design notes reference descriptions

100%

100%

Without context: $0.1729 · 55s · 12 turns · 16 in / 3,083 out tokens

With context: $0.4736 · 1m 57s · 23 turns · 186 in / 6,748 out tokens

100%

File Processing Agent Tools

Tool error handling and recovery

Criteria
Without context
With context

Explicit error returns

100%

100%

No silent failures

100%

100%

Error messages are descriptive

100%

100%

Recovery hints in errors

100%

100%

File-not-found handled

100%

100%

Schema validation errors informative

100%

100%

Transform errors specific

100%

100%

Success path uses consistent structure

100%

100%

Test file demonstrates error cases

100%

100%

No exception propagation to caller

100%

100%

Without context: $0.7110 · 2m 28s · 28 turns · 33 in / 11,339 out tokens

With context: $0.6335 · 2m 33s · 27 turns · 65 in / 10,367 out tokens

100%

Note-Taking MCP Server

MCP server tool design

Criteria
Without context
With context

Uses mcp package

100%

100%

MCP tool registration

100%

100%

Minimal tool count

100%

100%

Tool descriptions are substantive

100%

100%

Descriptions guide usage

100%

100%

Tool audit lists excluded tools

100%

100%

Tool audit explains descriptions

100%

100%

MCP server entry point

100%

100%

Parameter schemas defined

100%

100%

Error handling present

100%

100%

README documents tools

100%

100%

Without context: $0.2650 · 1m 21s · 18 turns · 22 in / 4,341 out tokens

With context: $0.6343 · 2m 42s · 31 turns · 95 in / 9,233 out tokens

Evaluated
Agent
Claude Code

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.