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.
npx skills add majiayu000/claude-skill-registry --skill agent-tool-builder-sickn33-antigravity-awesome --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Guides the creation of tools for AI agents, focusing on tool schema design, description quality, validation, and error handling. Stresses returning strings over objects and keeping tool counts under about twenty. Promotes testing tools with the LLM and using MCP standards for cross-platform tool design.
How it works
Outlines core capabilities and patterns for tool design:
- Emphasizes comprehensive, unambiguous tool descriptions as the primary driver of LLM accuracy.
- Recommends a small set of tools (fewer than 20) to reduce confusion.
- Requires explicit error handling to avoid silent failures.
- Advises returning strings rather than objects for LLM consumption.
- Mandates validation gates before execution to reject, fix, or escalate.
- Encourages testing tools with the LLM, not solely unit tests. Provides guidance on tool schema best practices, including detailed descriptions, parameter descriptions with format and examples, and explicit required vs optional fields. Includes usage examples for tool schemas and error handling, and introduces MCP tooling standards and tool runners for cross-provider compatibility.
When to use it
Triggered when defining or refining a new tool for an agent, or when establishing tool schemas and error-handling conventions. Also relevant when designing multi-tool coordination and cross-provider tool integrations.
What it can touch
Mentions standards and frameworks used for tool integration, including: JSON Schema, MCP (Model Context Protocol), Anthropic SDK, OpenAI Functions, Vercel AI SDK, LangChain Tools. Describes patterns for tool schema design, error handling, and tool use examples, but does not enumerate concrete touch points beyond these references.
Caveats
Risk is listed as unknown. License is MIT. The instruction set emphasizes that descriptions matter more than implementations and cautions against silent failures and vague descriptions. No explicit guarantees of performance or success are provided. The content centers on recommended practices and patterns without asserting outcomes.
# 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 implementations. The LLM never sees your code - it only sees the schema and description. ## Principles - Description quality > implementation quality for LLM accuracy - Aim for fewer than 20 tools - more causes confusion - Every tool needs explicit error handling - silent failures poison agents - Return strings, not objects - LLMs process text - Validation gates before execution - reject, fix, or escalate, never silent fail - Test tools with the LLM, not just unit tests ## Capabilities - agent-tools - function-calling - tool-schema-design - mcp-tools - tool-validation - tool-error-handling ## Scope - multi-agent-coordination → multi-agent-orchestration - agent-memory → agent-memory-
- Principles
- Capabilities
- Scope
- Tooling
- Standards
- Frameworks
- Patterns
- Tool Schema Design
- 1. Detailed Descriptions (Most Important)
- 2. Parameter Descriptions
- 3. Use Enums When Possible
- 4. Required vs Optional
- Tool with Input Examples
- Tool Error Handling
What does the agent-tool-builder skill do?
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.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill agent-tool-builder-sickn33-antigravity-awesome --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.