tool-design

This skill should be used when the user asks to "design agent tools", "create tool descriptions", "reduce tool complexity", "implement MCP tools", or mentions tool consolidation, architectural reduction, tool naming conventions, or agent-tool interfaces.

Install with Tessl CLI

npx tessl i github:muratcankoylan/Agent-Skills-for-Context-Engineering --skill tool-design

What are skills?

0.98x

Quality

41%

Does it follow best practices?

Impact

89%

0.98x

Average score across 3 eval scenarios

Optimize this skill with Tessl

npx tessl skill review --optimize ./skills/tool-design/SKILL.md

SKILL.md

Review

Evals

Evaluation results

86%

Improving Agent Tool Definitions

Tool description engineering

Criteria

Without context

With context

What the tool does

100%

When to use it

100%

Input documentation

100%

Return value documentation

100%

No vague descriptions

100%

Descriptive parameter names

100%

Concrete examples

100%

Meaningful defaults

62%

50%

Response format option

Actionable error documentation

60%

100%

Consistent naming

100%

Without context: $0.1769 · 55s · 7 turns · 8 in / 3,024 out tokens

With context: $0.3722 · 1m 26s · 14 turns · 15 in / 4,202 out tokens

82%

-13%

Redesigning a Scheduling Agent's Tool Set

Tool collection consolidation

Criteria

Without context

With context

Tool count reduced

100%

Sequential steps consolidated

92%

Date-range variants merged

100%

Update operations consolidated

100%

Namespacing used

70%

100%

Verb-noun naming

100%

70%

Unambiguous tool boundaries

100%

Consistent parameter naming

87%

Consolidation rationale

100%

Mutual availability generalized

100%

Without context: $0.2635 · 1m 17s · 8 turns · 9 in / 5,218 out tokens

With context: $0.4712 · 1m 51s · 16 turns · 14 in / 6,539 out tokens

100%

Agent System Prompt and Architecture Review for a Data Pipeline Assistant

MCP naming and architectural reduction

Criteria

Without context

With context

MCP fully qualified names

77%

100%

No unqualified MCP references

100%

All three servers addressed

100%

Reduction principle applied

100%

Wrapper tools challenged

100%

Documentation over tooling

100%

Model improvement argument

80%

100%

Maintenance burden acknowledged

100%

When to keep complexity

100%

Without context: $0.2625 · 1m 32s · 11 turns · 11 in / 4,560 out tokens

With context: $0.4139 · 1m 53s · 15 turns · 14 in / 4,877 out tokens

Evaluated: 5 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Improving Agent Tool Definitions Redesigning a Scheduling Agent's Tool Set Agent System Prompt and Architecture Review for a Data Pipeline Assistant

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.