api-inferrer-from-files

What this skill does

# API Endpoint Inferrer

Generate a concise list of REST API endpoints inferred from file names or project structures. Output is intentionally minimal: one line per endpoint — the method + path + a single-sentence description. No schemas, no request bodies, no examples, no documentation blocks.

---

## Input Formats Accepted

- A flat list of file names: `userController.js`, `productService.py`, etc.
- A directory tree (from `tree`, `ls -R`, `find`, or copy-pasted folder structure)
- A mix of both
- File names with or without extensions
- Any backend language/framework (Node, Python, Go, Ruby, Java, PHP, etc.)

---

## Inference Rules

### 1. Extract Resources
Identify the core resource name from each file. Strip common suffixes:
- `Controller`, `Router`, `Route`, `Handler`, `Service`, `Model`, `View`, `Schema`, `Repository`, `Repo`, `Store`, `Manager`, `API`

Examples:
- `userController.js` → resource: `user`
- `product_routes.py` → resource: `product`
- `OrderService.java` → resource: `order`
- `auth.go` → resource: `auth`

### 2. Map Files to Standard CRUD Endpoints
For each resource, infer standard REST endpoints based on typical conventions:

| Pattern | Endpoints to infer |
|---|---|
| `*Controller`, `*Router`, `*Route`, `*Handler` | Full CRUD: GET list, GET by ID, POST, PUT/PATCH, DELETE |
| `*Service` | Same as controller (services back controllers) |
| `*Model`, `*Schema` | GET list, GET by ID, POST (data-centric) |
| `auth*`, `*Auth*` | POST /login, POST /logout, POST /register, POST /refresh |
| `*upload*`, `*file*`, `*media*` | POST upload, GET file by ID, DELETE file |
| `*search*` | GET /search with query params |
| `*config*`, `*settings*` | GET settings, PUT settings |
| `*health*`, `*status*`, `*ping*` | GET /health or GET /status |
| `*webhook*` | POST /webhook |
| `*middleware*`, `*interceptor*` | Skip — not an endpoint |
| `*util*`, `*helper*`, `*constant*` | Skip — not an endpoint |
| `*test*`, `*spec*`, `*mock*` | Skip — not an endpoint |

### 3. Naming Convention → URL Path
Convert resource names to kebab-case plural paths:
- `User` → `/users`
- `ProductCategory` → `/product-categories`
- `OrderItem` → `/order-items`
- `auth` stays → `/auth`

### 4. Version Prefix (Optional)
If project structure contains a `v1/`, `v2/`, `api/` folder — prefix routes accordingly (e.g., `/api/v1/users`). Otherwise use bare paths.

### 5. Nested Resources
If two related files exist (e.g., `commentController` alongside `postController`), infer nested routes:
- `GET /posts/:id/comments`
- `POST /posts/:id/comments`

---

## Output Format

Output a clean, unstyled list. No headers, no bold, no code blocks unless the user asks. Just:

```
METHOD /path/to/endpoint — One sentence explaining what it does.
```

Group by resource. Separate groups with a blank line. Keep descriptions under 12 words.

### Example Output

Given files: `userController.js`, `productRouter.js`, `authService.js`, `orderModel.js`

---

GET /users — Retrieve a list of all users.
GET /users/:id — Fetch a single user by ID.
POST /users — Create a new user.
PUT /users/:id — Update an existing user's details.
DELETE /users/:id — Remove a user by ID.

GET /products — List all available products.
GET /products/:id — Get details of a specific product.
POST /products — Add a new product.
PUT /products/:id — Update a product's information.
DELETE /products/:id — Delete a product.

POST /auth/register — Register a new account.
POST /auth/login — Authenticate and receive a token.
POST /auth/logout — Invalidate the current session.
POST /auth/refresh — Refresh an expired access token.

GET /orders — List all orders.
GET /orders/:id — Retrieve a specific order.
POST /orders — Place a new order.

---

## Edge Cases

- **Unknown/ambiguous file names** (e.g., `utils.js`, `index.ts`): Skip unless context makes the resource clear.
- **Multiple files for same resource** (e.g., `userController.js` + `userRouter.js`): Deduplicate — list each endpoint only once.
- **Non-REST files** (e.g., `app.js`, `server.py`, `main.go`, `config.json`): Skip.
- **GraphQL hint** (e.g., `schema.graphql`, `resolvers.js`): Mention that this appears to be a GraphQL API and list query/mutation names instead of REST paths.
- **If no backend files detected**: Respond with "No API-related files detected. Please share controller, route, service, or model file names."

---

## Tone

- Be terse. The user explicitly wants brevity.
- Do not explain your reasoning unless asked.
- Do not add notes like "This assumes REST conventions" unless something genuinely unusual was inferred.
- If input is ambiguous, make a best-guess and list it — don't ask clarifying questions unless the file names give literally no signal.


---

## After Completing the API Output

Once the API output is delivered, ask the user:

"Would you like me to design the APIs for these endpoints? (yes/no)"

If the user says **yes**:
- Check if the API Designer skill is available in the installed skills list
- If the skill **is available**:
  - Read and follow the instructions in the API Designer skill
  - Use the API design output above as the input
- If the skill **is NOT available**:
  - Inform the user: "It looks like the API Designer skill isn't installed. 
    You can install it and re-run.

If the user says **no**:
- End the task here

---