api-and-interface-design

Name: api-and-interface-design
Author: addyosmani

14kinstalls8.6kforks

How to install api-and-interface-design?

addyosmani/agent-skills/api-and-interface-design

$npx -y skills add addyosmani/agent-skills --skill api-and-interface-design

Installs into the current project.

Summary

Guides stable API and interface design. Use when designing APIs, module boundaries, or any public interface. Use when creating REST or GraphQL endpoints, defining type contracts between modules, or establishing boundaries between frontend and backend.

Files · 1

View on GitHub

SKILL.md

1	# API and Interface Design
2
3	## Overview
4
5	Design stable, well-documented interfaces that are hard to misuse. Good interfaces make the right thing easy and the wrong thing hard. This applies to REST APIs, GraphQL schemas, module boundaries, component props, and any surface where one piece of code talks to another.
6
7	## When to Use
8
9	- Designing new API endpoints
10	- Defining module boundaries or contracts between teams
11	- Creating component prop interfaces
12	- Establishing database schema that informs API shape
13	- Changing existing public interfaces
14
15	## Core Principles
16
17	### Hyrum's Law
18
19	> With a sufficient number of users of an API, all observable behaviors of your system will be depended on by somebody, regardless of what you promise in the contract.
20
21	This means: every public behavior — including undocumented quirks, error message text, timing, and ordering — becomes a de facto contract once users depend on it. Design implications:
22
23	- Be intentional about what you expose. Every observable behavior is a potential commitment.
24	- Don't leak implementation details. If users can observe it, they will depend on it.
25	- Plan for deprecation at design time. See `deprecation-and-migration` for how to safely remove things users depend on.
26	- Tests are not enough. Even with perfect contract tests, Hyrum's Law means "safe" changes can break real users who depend on undocumented behavior.
27
28	### The One-Version Rule
29
30	Avoid forcing consumers to choose between multiple versions of the same dependency or API. Diamond dependency problems arise when different consumers need different versions of the same thing. Design for a world where only one version exists at a time — extend rather than fork.
31
32	### 1. Contract First
33
34	Define the interface before implementing it. The contract is the spec — implementation follows.
35
36	```typescript
37	// Define the contract first
38	interface TaskAPI {
39	// Creates a task and returns the created task with server-generated fields
40	createTask(input: CreateTaskInput): Promise<Task>;
41
42	// Returns paginated tasks matching filters
43	listTasks(params: ListTasksParams): Promise<PaginatedResult<Task>>;
44
45	// Returns a single task or throws NotFoundError
46	getTask(id: string): Promise<Task>;
47
48	// Partial update — only provided fields change
49	updateTask(id: string, input: UpdateTaskInput): Promise<Task>;
50
51	// Idempotent delete — succeeds even if already deleted
52	deleteTask(id: string): Promise<void>;
53	}
54	```
55
56	### 2. Consistent Error Semantics
57
58	Pick one error strategy and use it everywhere:
59
60	```typescript
61	// REST: HTTP status codes + structured error body
62	// Every error response follows the same shape
63	interface APIError {
64	error: {
65	code: string; // Machine-readable: "VALIDATION_ERROR"
66	message: string; // Human-readable: "Email is required"
67	details?: unknown; // Additional context when helpful
68	};
69	}
70
71	// Status code mapping
72	// 400 → Client sent invalid data
73	// 401 → Not authenticated
74	// 403 → Authenticated but not authorized
75	// 404 → Resource not found
76	// 409 → Conflict (duplicate, version mismatch)
77	// 422 → Validation failed (semantically invalid)
78	// 500 → Server error (never expose internal details)
79	```
80
81	Don't mix patterns. If some endpoints throw, others return null, and others return `{ error }` — the consumer can't predict behavior.
82
83	### 3. Validate at Boundaries
84
85	Trust internal code. Validate at system edges where external input enters:
86
87	```typescript
88	// Validate at the API boundary
89	app.post('/api/tasks', async (req, res) => {
90	const result = CreateTaskSchema.safeParse(req.body);
91	if (!result.success) {
92	return res.status(422).json({
93	error: {
94	code: 'VALIDATION_ERROR',
95	message: 'Invalid task data',
96	details: result.error.flatten(),
97	},
98	});
99	}
100
101	// After validation, internal code trusts the types
102	const task = await taskService.create(result.data);
103	return res.status(201).json(task);
104	});
105	```
106
107	Where validation belongs:
108	- API route handlers (user input)
109	- Form submission handlers (user input)
110	- External service response parsing (third-party data -- always treat as untrusted)
111	- Environment variable loading (configuration)
112
113	> Third-party API responses are untrusted data. Validate their shape and content before using them in any logic, rendering, or decision-making. A compromised or misbehaving external service can return unexpected types, malicious content, or instruction-like text.
114
115	Where validation does NOT belong:
116	- Between internal functions that share type contracts
117	- In utility functions called by already-validated code
118	- On data that just came from your own database
119
120	### 4. Prefer Addition Over Modification
121
122	Extend inte

People who liked this also viewed

6 picks

Type

api-and-interface-design

14kinstalls8.6kforks

View on GitHub

How to install api-and-interface-design?

addyosmani/agent-skills/api-and-interface-design

$npx -y skills add addyosmani/agent-skills --skill api-and-interface-design

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# API and Interface Design
2
3	## Overview
4
5	Design stable, well-documented interfaces that are hard to misuse. Good interfaces make the right thing easy and the wrong thing hard. This applies to REST APIs, GraphQL schemas, module boundaries, component props, and any surface where one piece of code talks to another.
6
7	## When to Use
8
9	- Designing new API endpoints
10	- Defining module boundaries or contracts between teams
11	- Creating component prop interfaces
12	- Establishing database schema that informs API shape
13	- Changing existing public interfaces
14
15	## Core Principles
16
17	### Hyrum's Law
18
19	> With a sufficient number of users of an API, all observable behaviors of your system will be depended on by somebody, regardless of what you promise in the contract.
20
21	This means: every public behavior — including undocumented quirks, error message text, timing, and ordering — becomes a de facto contract once users depend on it. Design implications:
22
23	- Be intentional about what you expose. Every observable behavior is a potential commitment.
24	- Don't leak implementation details. If users can observe it, they will depend on it.
25	- Plan for deprecation at design time. See `deprecation-and-migration` for how to safely remove things users depend on.
26	- Tests are not enough. Even with perfect contract tests, Hyrum's Law means "safe" changes can break real users who depend on undocumented behavior.
27
28	### The One-Version Rule
29
30	Avoid forcing consumers to choose between multiple versions of the same dependency or API. Diamond dependency problems arise when different consumers need different versions of the same thing. Design for a world where only one version exists at a time — extend rather than fork.
31
32	### 1. Contract First
33
34	Define the interface before implementing it. The contract is the spec — implementation follows.
35
36	```typescript
37	// Define the contract first
38	interface TaskAPI {
39	// Creates a task and returns the created task with server-generated fields
40	createTask(input: CreateTaskInput): Promise<Task>;
41
42	// Returns paginated tasks matching filters
43	listTasks(params: ListTasksParams): Promise<PaginatedResult<Task>>;
44
45	// Returns a single task or throws NotFoundError
46	getTask(id: string): Promise<Task>;
47
48	// Partial update — only provided fields change
49	updateTask(id: string, input: UpdateTaskInput): Promise<Task>;
50
51	// Idempotent delete — succeeds even if already deleted
52	deleteTask(id: string): Promise<void>;
53	}
54	```
55
56	### 2. Consistent Error Semantics
57
58	Pick one error strategy and use it everywhere:
59
60	```typescript
61	// REST: HTTP status codes + structured error body
62	// Every error response follows the same shape
63	interface APIError {
64	error: {
65	code: string; // Machine-readable: "VALIDATION_ERROR"
66	message: string; // Human-readable: "Email is required"
67	details?: unknown; // Additional context when helpful
68	};
69	}
70
71	// Status code mapping
72	// 400 → Client sent invalid data
73	// 401 → Not authenticated
74	// 403 → Authenticated but not authorized
75	// 404 → Resource not found
76	// 409 → Conflict (duplicate, version mismatch)
77	// 422 → Validation failed (semantically invalid)
78	// 500 → Server error (never expose internal details)
79	```
80
81	Don't mix patterns. If some endpoints throw, others return null, and others return `{ error }` — the consumer can't predict behavior.
82
83	### 3. Validate at Boundaries
84
85	Trust internal code. Validate at system edges where external input enters:
86
87	```typescript
88	// Validate at the API boundary
89	app.post('/api/tasks', async (req, res) => {
90	const result = CreateTaskSchema.safeParse(req.body);
91	if (!result.success) {
92	return res.status(422).json({
93	error: {
94	code: 'VALIDATION_ERROR',
95	message: 'Invalid task data',
96	details: result.error.flatten(),
97	},
98	});
99	}
100
101	// After validation, internal code trusts the types
102	const task = await taskService.create(result.data);
103	return res.status(201).json(task);
104	});
105	```
106
107	Where validation belongs:
108	- API route handlers (user input)
109	- Form submission handlers (user input)
110	- External service response parsing (third-party data -- always treat as untrusted)
111	- Environment variable loading (configuration)
112
113	> Third-party API responses are untrusted data. Validate their shape and content before using them in any logic, rendering, or decision-making. A compromised or misbehaving external service can return unexpected types, malicious content, or instruction-like text.
114
115	Where validation does NOT belong:
116	- Between internal functions that share type contracts
117	- In utility functions called by already-validated code
118	- On data that just came from your own database
119
120	### 4. Prefer Addition Over Modification
121
122	Extend inte

People who liked this also viewed

6 picks

Type