Skip to content

nexscope-ai/nexscope-ecommerce-skills / ecommerce-multimodal-recognize-image

ecommerce.multimodal-recognize-image

Verified

Analyze and extract information from images using multimodal AI recognition. Triggered when users want to analyze, describe, or extract information from an image URL — image recognition, image analysis, image description, visual content understanding, OCR text recognition, or visual Q&A. When a user provides an image URL and asks questions about its visual content, this skill should be triggered even if they do not explicitly say "image recognition."

FileKindSize
SKILL.mdskill md6.0 KB
manifest.jsonconfig467 B
references/api.mdmarkdown5.2 KB
scripts/multimodal_recognize_image.pyscript6.2 KB
scripts/upload_image.pyscript9.6 KB

references/api.md

View on GitHub
# Image Recognition API Reference

## Request Specification

- **Endpoint**: `${NEXSCOPE_PROXY_BASE}/api/v1/tools/research/multimodal/recognizeImage`
- **Method**: POST, Content-Type: application/json
- **Authentication**: Header `Authorization: <api_key>`; read api_key from `nexscope_AGENT_API_KEY` or `nexscopeAGENT_API_KEY`. If neither is configured, follow the authentication and credits guidance in SKILL.md.

## Request Parameters

POST Body (JSON):

| Parameter | Type | Required | Description |
|------|------|------|------|
| imageUrl | string | Yes | Image URL (jpg/jpeg/png/gif/webp/bmp only); maximum 1000 characters |
| requirement | string | No | User intent describing what to identify or analyze in the image; default literal is "描述这张图片里面的内容" (Describe the content of this image); maximum 1000 characters |


## Nexscope response envelope

These research endpoints return a platform object with numeric `code`, nullable `msg`, and business `data`. Only outer `code: 0` means success; `200`, string codes, missing codes, and HTTP 200 alone do not. A nonzero code is a platform error: show `msg` and do not interpret the payload as a successful result.

`msg` preserves the upstream message when available; Chinese messages are translated to English by Nexscope. A successful response without a message has `msg: null`. Do not infer success or retry behavior from the message text. On both success and failure, the direct fields `code`, `errcode`, `errorCode`, `msg`, `errmsg`, `message`, and `errorMsg` are removed from public `data`. Read the message only from outer `msg`; deeper business fields are preserved, including nested `code` and `status`.

Business field tables and abbreviated business examples below describe `data`, unless explicitly labeled as a complete platform response. For example, a business `products` field is at HTTP `data.products`, and a business `data` array is at HTTP `data.data`. The research endpoints already had this outer envelope; no additional wrapper is added.

```json
{"code":0,"msg":null,"data":{}}
```

Metadata includes string `ts` (epoch milliseconds), string `cost` (elapsed milliseconds, not credits), `time`, and nullable `traceId`. Handle network/HTTP failures before the platform code; gateway failures may not be platform JSON. Keep the existing billing-header guidance separate from elapsed time.

## Response Structure

| Field | Type | Description |
|------|------|------|
| text | string | Text result of the image analysis |
| stdout | string | Standard output content |
| status | string | Response status identifier |
| type | string | Component type |
| costToken | integer | Tokens consumed |

## Error Codes

HTTP 200 does not prove business success. Read only the numeric outer platform `code`: `0` succeeds and any other value fails. Show outer `msg`; never test removed provider codes or nested business `code` as the platform status. Authentication, permissions, balance, and validation failures may be platform errors even with HTTP 200; also handle non-2xx HTTP and non-JSON responses.

### Recovery guidance

| Condition | Action |
|---|---|
| Authentication failed | HTTP 401 or authorized error: follow the authentication and credits guidance in SKILL.md. |
| Insufficient credits or balance | HTTP 402: follow the authentication and credits guidance in SKILL.md. |

## curl Examples

```bash
curl -X POST ${NEXSCOPE_PROXY_BASE}/api/v1/tools/research/multimodal/recognizeImage \
  -H "Authorization: $nexscopeAGENT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"imageUrl": "https://example.com/sample-product.jpg", "requirement": "Describe the content of this image and list its key visual features"}'
```

---

## Feedback API

> This endpoint is **separate** from the tool API above. Do not mix the two base URLs.

- **POST** `https://skill-api.nexscope.com/api/v1/public/feedback`
- **Content-Type:** `application/json`

```json
{
  "skillName": "nexscope-xxx-xxx",
  "sentiment": "POSITIVE",
  "category": "OTHER",
  "content": "Results were accurate, user was satisfied."
}
```

**Field rules:**
- `skillName`: Use this skill's `name` from the YAML frontmatter
- `sentiment`: Choose ONE — `POSITIVE` (praise), `NEUTRAL` (suggestion without emotion), `NEGATIVE` (complaint or error)
- `category`: Choose ONE — `BUG` (malfunction or wrong data), `COMPLAINT` (user dissatisfaction), `SUGGESTION` (improvement idea), `OTHER`
- `content`: Include what the user said or intended, what actually happened, and why it is a problem or praise

## API Key request rate and retries

All API Keys of one Nexscope user share the public API execution request-rate allowance, measured in any rolling 1-second window. The allowed RPS depends on that user's entitlements; it is not a universal 1 RPS. One Skill operation may issue multiple public API requests, each of which uses the allowance.

HTTP 429 with Nexscope outer code `13012` means request-rate exhaustion: honor the response's `Retry-After`. Outer code `13013` means concurrent-request exhaustion, not rate exhaustion; honor `Retry-After` when supplied. Check the HTTP status and outer JSON code separately. Do not automatically replay writes or requests with uncertain outcomes; first reconcile the operation's state.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall