TencentCloud General OCR

腾讯云广告文字识别(AdvertiseOCR)接口调用技能。当用户需要从图片中识别文字内容时,应使用此技能。支持中英文、横排、竖排及倾斜场景的图片文字识别,支持90度、180度、270度翻转场景的图片识别,返回文本框位置与文字内容。支持图片Base64和URL两种输入方式。

Security Scan

The skill appears to implement Tencent Cloud AdvertiseOCR correctly, but the registry metadata omits required Tencent Cloud credentials and there is no automated install spec — this mismatch is concerning and should be resolved before trusting the skill.

Purpose & Capability

The code and SKILL.md implement a Tencent Cloud AdvertiseOCR client and legitimately require Tencent Cloud API keys; the requested capabilities (OCR via tencentcloud-sdk-python) align with the skill's stated purpose. However, the registry metadata claims no required environment variables/credentials while both the instructions and script require TENCENTCLOUD_SECRET_ID and TENCENTCLOUD_SECRET_KEY — this metadata omission is inconsistent and reduces trust.

Instruction Scope

The runtime instructions and script stay within OCR scope: they accept an image URL or base64/file, read files only when explicitly provided, call the official OCR endpoint (ocr.tencentcloudapi.com), and format the response. The script reads environment variables for credentials and does not access unrelated system paths or exfiltrate data to unexpected endpoints.

Install Mechanism

There is no install spec in the registry (instruction-only), but SKILL.md and the script require the third-party package tencentcloud-sdk-python and instruct the user to pip install it. Lack of an explicit install entry means dependencies won't be auto-installed and package versions are unpinned — this is reasonable for an instruction-only skill but raises supply-chain and reproducibility concerns.

Credentials

The only sensitive items the skill needs are Tencent Cloud API keys (TENCENTCLOUD_SECRET_ID and TENCENTCLOUD_SECRET_KEY), which are proportionate to calling the Tencent OCR API. The concern is the inconsistency: the registry metadata lists no required env vars while the SKILL.md and script require secrets; that mismatch could lead to accidental credential exposure or user confusion about what will be requested.

Persistence & Privilege

The skill does not request permanent presence, does not set always:true, and does not modify other skills or system-wide configurations. It runs as a standalone script and prints results to stdout.

What to consider before installing

This skill's code and documentation implement Tencent Cloud AdvertiseOCR and require your Tencent Cloud API keys (TENCENTCLOUD_SECRET_ID and TENCENTCLOUD_SECRET_KEY), but the registry metadata does not declare those env vars — treat that as a red flag. Before installing or running: (1) verify the source and trustworthiness of the skill owner; (2) inspect the included scripts (you already have main.py) and confirm there are no unexpected network endpoints; (3) provide credentials only in a controlled environment (use short-lived keys or a restricted account if possible); (4) run the script in an isolated environment (container) and avoid passing files that contain other secrets; (5) consider adding an explicit install step that pins tencentcloud-sdk-python to a known-safe version, or manually install dependencies after review. If you plan to enable this skill for autonomous agent use, resolve the metadata inconsistency first so users understand what credentials will be required.

腾讯云广告文字识别 (AdvertiseOCR)

用途

调用腾讯云OCR广告文字识别接口，支持图片内文字的检测和识别，返回文本框位置与文字内容。具有较高召回率和准确率。

核心能力：

• 中英文识别：支持中英文混合文字识别
• 多方向支持：支持横排、竖排以及倾斜场景文字识别
• 翻转支持：支持90度、180度、270度翻转场景文字识别
• 坐标返回：返回每个文本行的四顶点坐标（Polygon）
• 置信度评估：返回每个文本行的识别置信度（0~100）

官方文档：https://cloud.tencent.com/document/api/866/49524

默认接口请求频率限制：20次/秒。

使用时机

当用户提出以下需求时触发此技能：

• 需要从图片中提取文字信息
• 需要识别图片上的文字内容
• 涉及文字OCR识别的任何场景
• 需要获取图片中文字的位置坐标信息

环境要求

• Python 3.6+
• 依赖：tencentcloud-sdk-python（通过 pip install tencentcloud-sdk-python 安装）
• 环境变量：
  • TENCENTCLOUD_SECRET_ID：腾讯云API密钥ID
  • TENCENTCLOUD_SECRET_KEY：腾讯云API密钥Key

使用方式

运行 scripts/main.py 脚本完成图片文字识别。

请求参数

参数	类型	必填	说明
ImageBase64	str	否(二选一)	图片Base64值，编码后不超过10M，分辨率建议600*800以上，支持PNG/JPG/JPEG/BMP
ImageUrl	str	否(二选一)	图片URL地址，建议存储于腾讯云COS。都提供时仅使用ImageUrl
UserAgent	str	否	请求来源标识(可选)，用于追踪调用来源，统一固定为Skills

输出格式

识别成功后返回 JSON 格式结果：

{
  "TextDetections": [
    {
      "DetectedText": "识别出的文本行内容",
      "Confidence": 99,
      "Polygon": [
        {"X": 0, "Y": 0},
        {"X": 100, "Y": 0},
        {"X": 100, "Y": 50},
        {"X": 0, "Y": 50}
      ],
      "AdvancedInfo": "{\"Parag\":{\"ParagNo\":1}}"
    }
  ],
  "TextCount": 1,
  "ImageSize": {
    "Width": 800,
    "Height": 600
  },
  "RequestId": "xxx"
}

响应字段说明：

字段	类型	说明
TextDetections	list	检测到的文本信息列表
TextDetections[].DetectedText	str	识别出的文本行内容
TextDetections[].Confidence	int	置信度 0~100
TextDetections[].Polygon	list of Coord	文本行坐标，四个顶点坐标（X, Y）
TextDetections[].AdvancedInfo	str	扩展字段，含段落信息Parag(ParagNo)
TextCount	int	检测到的文本行数量
ImageSize	object	图片分辨率信息，含Width和Height（单位px）
RequestId	str	唯一请求ID

错误码说明

错误码	含义
FailedOperation.DownLoadError	文件下载失败
FailedOperation.EmptyImageError	图片内容为空
FailedOperation.EngineRecognizeTimeout	引擎识别超时
FailedOperation.ImageDecodeFailed	图片解码失败
FailedOperation.ImageNoText	图片中未检测到文本
FailedOperation.LanguageNotSupport	输入的Language不支持
FailedOperation.OcrFailed	OCR识别失败
FailedOperation.UnKnowError	未知错误
FailedOperation.UnOpenError	服务未开通
InvalidParameterValue.InvalidParameterValueLimit	参数值错误
LimitExceeded.TooLargeFileError	文件内容太大
ResourceUnavailable.InArrears	账号已欠费
ResourceUnavailable.ResourcePackageRunOut	账号资源包耗尽
ResourcesSoldOut.ChargeStatusException	计费状态异常

业务逻辑说明

1. ImageBase64和ImageUrl必须提供其一，都提供时只使用ImageUrl
2. 图片经Base64编码后不超过10M，分辨率建议600*800以上
3. 支持PNG、JPG、JPEG、BMP格式
4. 始终计费

调用示例

# 通过URL识别图片文字
python scripts/main.py --image-url "https://example.com/ad_image.jpg"

# 通过文件路径(自动Base64编码)识别
python scripts/main.py --image-base64 ./ad_image.jpg

# 通过Base64文本文件识别
python scripts/main.py --image-base64 ./base64.txt

# 指定地域
python scripts/main.py --image-url "https://example.com/ad_image.jpg" --region ap-beijing```