JPeetz/agent-skills / skills/development/finding-unknowns/skills/agent-interface-design
agent-interface-design
VerifiedUse when building MCP servers or agent-facing tools.
| File | Kind | Size |
|---|---|---|
| CHANGELOG.md | markdown | 405 B |
| SKILL.md | skill md | 2.4 KB |
| evals/evals.json | config | 2.5 KB |
evals/evals.json
View on GitHub{
"skill": "agent-interface-design",
"version": "1.0.0",
"cases": [
{
"id": "e1",
"description": "Should trigger: building MCP servers or agent-facing tools",
"input": "Design the MCP server interface for our internal search tool.",
"shouldTrigger": true,
"expect": "Designs agent-facing interface: enumerated params, intent-named fields, description carries behavior",
"assertions": [
"Pushes meaning into parameters (enumerate instead of free text)",
"Names for intent not implementation",
"Makes invalid states unrepresentable where type system allows",
"Puts behavioral instruction in the tool's own description",
"Names the mistake the design still permits and whether to guard"
]
},
{
"id": "e2",
"description": "Should trigger: misuse-diagnosed tool redesign",
"input": "Agents keep calling our tool wrong even with docs. Redesign the interface.",
"shouldTrigger": true,
"expect": "Treats misuse as interface symptom first, fixes with rename/type before prose",
"assertions": [
"Reads transcripts/logs of misuse first",
"Fixes interface before adding documentation",
"Keeps an example only for unguessable formats"
]
},
{
"id": "n1",
"description": "Not trigger: implementing an existing interface, not designing one",
"input": "Implement the REST endpoints for the search tool as specced.",
"shouldTrigger": false,
"expect": "Implements per spec",
"assertions": [
"No interface redesign"
]
},
{
"id": "n2",
"description": "Not trigger: human-facing UI design",
"input": "Design a dashboard page for our ops team with charts.",
"shouldTrigger": false,
"expect": "Treats as human UI work",
"assertions": [
"No agent interface machinery (MCP/tool descriptions) involved"
]
},
{
"id": "e3",
"description": "Should trigger: locating behavioral guidance at point of use",
"input": "We need to tell agents not to retry this tool more than twice. Where should that live?",
"shouldTrigger": true,
"expect": "Puts behavioral instruction in the tool's own description at the point of use, not a global preamble",
"assertions": [
"Places the constraint in the tool's own description",
"Does not add it to a global preamble",
"Checks misuse transcripts first if redesign is involved"
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |