← Back to Skills Marketplace
TencentCloud VITA
by
Wayne-J151
· GitHub ↗
· v1.0.9
· MIT-0
504
Downloads
0
Stars
0
Active Installs
8
Versions
Install in OpenClaw
/install tencentcloud-vita
Description
腾讯云图像与视频理解 (TencentCloud VITA / vita) 技能。用于图片理解、视频理解、图像分析、视频分析、描述生成、监控录像分析、体育高光提取、直播理解、新闻视频摘要、产品演示视频分析，以及设置或更新 vita prompt。支持单张图片、多张图片、视频。TencentCloud VITA (...
Usage Guidance
This skill appears to implement TencentCloud VITA integration, and the bundled script is coherent with the SKILL.md. However: (1) the registry metadata omitted the required environment variable TENCENTCLOUD_VITA_API_KEY — make sure you set that only if you trust this skill and its source; (2) the script requires the 'openai' Python package (pip install openai); (3) the skill will read user-specified local image files and base64-encode them for upload, and it will write a prompt file to <SKILL_DIR>/prompt/vita_prompt.txt when asked to persist a prompt. Before installing, confirm the API endpoint (VITA_BASE_URL) is the official Tencent endpoint you expect and prefer the prompt persistence behavior. If you need higher confidence, ask the publisher to update the registry metadata to declare the required env var and provide a homepage/source validation.
Capability Analysis
Type: OpenClaw Skill
Name: tencentcloud-vita
Version: 1.0.9
The tencentcloud-vita skill provides a legitimate interface for Tencent Cloud's VITA image and video understanding service. The core logic in `scripts/main.py` uses the OpenAI SDK to interact with the official Tencent Cloud API endpoint and includes standard handling for local image files (base64 encoding). The instructions in `SKILL.md` regarding the persistence of a custom prompt in a local file are clearly documented as a feature for the agent to manage, and no evidence of malicious execution, data exfiltration, or harmful prompt injection was found.
Capability Assessment
Purpose & Capability
The SKILL.md and the bundled script clearly require a TencentCloud VITA API key (TENCENTCLOUD_VITA_API_KEY) and the openai Python SDK; however the registry metadata lists no required environment variables or primary credential. That omission is an incoherence between what the skill says it needs and what it actually uses. The code's network target (VITA_BASE_URL) and features (image/video analysis, base64 local-image encoding) are consistent with the described purpose.
Instruction Scope
The SKILL.md prescribes running the included CLI (scripts/main.py) and defines how to handle local images (read and base64-encode) and that local video upload must be done via a separate tool. The only non-trivial agent-side action mandated outside running the script is persisting a prompt to <SKILL_DIR>/prompt/vita_prompt.txt; this is scoped to the skill directory and consistent with the stated prompt-persistence feature. The instructions do not ask the agent to read unrelated system files or exfiltrate data beyond calls to the VITA API.
Install Mechanism
There is no install spec (instruction-only install), but the shipped Python script depends on the 'openai' package and will exit with a DEPENDENCY_MISSING message if it's not installed. That is reasonable but means manual pip install is required. No downloads from arbitrary URLs or extracted archives are present.
Credentials
The code requires a single credential: TENCENTCLOUD_VITA_API_KEY (read from the environment). That is proportional to the skill's function, but the registry metadata failing to declare this required env var is a discrepancy. No other secrets or unrelated environment variables are accessed by the script.
Persistence & Privilege
The skill persists a user-provided 'vita_prompt.txt' under its own prompt/ directory and instructs the agent to write that file when the user requests. Persisting configuration inside the skill's directory is expected behavior, but users should be aware the skill will store the prompt on disk. The skill is not marked 'always: true' and does not request broader system-wide privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install tencentcloud-vita - After installation, invoke the skill by name or use
/tencentcloud-vita - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.9
TencentCloud VITA 1.0.9 — documentation update
- Updated and improved SKILL documentation for clarity and consistency.
- Enhanced English descriptions; clarified skill’s name as TencentCloud VITA (vita) and its positioning for video/image understanding.
- Simplified language; restructured guide sections for easier agent integration and prompt management.
- Standardized environment variable to TENCENTCLOUD_VITA_API_KEY.
- No code or functional changes to the skill itself.
v1.0.7
腾讯云VITA视觉理解技能 v1.0.7 更新说明：
- 新增支持本地图片路径输入，脚本自动将本地图片转为 base64 data URL 后调用 VITA API。
- 明确说明本地视频不再由脚本内置上传，需额外借助 COS 等上传工具获取可访问 URL 后再分析。
- 调整参数说明与调用示例，标注 `--image` 支持 URL 或本地路径，`--video` 仅支持远程 URL。
- 优化文档对本地文件、上传流程的指引，避免误解。
- 增强失败约束：脚本不负责本地视频上传，遇本地视频须明确提示，不再自动引导或处理。
- 依赖与功能保持与上版一致，兼容原有 API KEY 和 PROMPT 配置方式。
v1.0.6
- 支持直接处理本地图片/视频文件：本地文件会自动上传到腾讯云 COS 并生成预签名 URL，无需用户手动上传。
- 增加“持久化 prompt”功能：允许用户设置/更新 VITA 视频理解提示词，保存为本地文件并在后续调用中使用。
- 新增详细指引，指导如何处理本地文件与 prompt 的优先级关系。
- 移除用户交互执行确认环节，执行行为简化。
- 描述及使用场景说明优化，明确“设置/修改 VITA prompt”场合也可调用此技能。
v1.0.4
tencentcloud-vita 1.0.4
- No code or documentation changes in this release.
- Functionality, usage instructions, and environment requirements remain the same as the previous version.
v1.0.3
- 初始发布，新增 tencentcloud-vita AI 视觉分析技能。
- 新增核心脚本 scripts/main.py，实现图像/视频理解与分析。
- 提供对单图、多图和视频 URL 输入的支持，可自定义分析 prompt。
- 支持流式与非流式输出，多种场景和详细调用示例。
- 包含详细 Agent 使用规范与安全说明，便于集成和部署。
v1.0.2
- Added explicit specification for the VITA_API_KEY environment variable in the skill manifest.
- Updated Agent execution protocol: Agent must now confirm the media URL, prompt, and API fee notice with the user before running analysis, and wait for user consent.
- Clarified that running the skill may incur costs by calling an external API.
- Changed dependency installation instructions to require manual installation of the openai SDK before running.
- Updated the list of agent behaviors to avoid, emphasizing the need for user confirmation before execution.
v1.0.0
TencentCloud VITA v1.0.0
- 初始发布，支持腾讯云VITA图像/视频理解AI服务。
- 支持多图片/单图片/单视频URL输入，以及自定义分析prompt。
- 提供详细的环境变量配置和API KEY获取说明。
- 完整展示命令行使用方式、参数含义与场景化prompt模板推荐。
- 明确列出Agent执行规范及须避免行为，保障分析结果准确性。
- 计费、依赖及并发限制信息透明展示。
v1.0.1
- Added comprehensive SKILL.md with detailed usage instructions, environment setup, parameter list, output formats, and recommended prompts.
- Clarified agent execution rules and strict requirements for handling user inputs, including error handling.
- Provided clear input format and media type/size limitations for image and video analysis.
- Listed cost details, script entry points, and dependency setup.
- Included best practice prompt templates for various use cases such as surveillance analysis, news summarization, e-commerce, sports highlights, and stream quality evaluation.
Metadata
Frequently Asked Questions
What is TencentCloud VITA?
腾讯云图像与视频理解 (TencentCloud VITA / vita) 技能。用于图片理解、视频理解、图像分析、视频分析、描述生成、监控录像分析、体育高光提取、直播理解、新闻视频摘要、产品演示视频分析，以及设置或更新 vita prompt。支持单张图片、多张图片、视频。TencentCloud VITA (... It is an AI Agent Skill for Claude Code / OpenClaw, with 504 downloads so far.
How do I install TencentCloud VITA?
Run "/install tencentcloud-vita" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is TencentCloud VITA free?
Yes, TencentCloud VITA is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does TencentCloud VITA support?
TencentCloud VITA is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created TencentCloud VITA?
It is built and maintained by Wayne-J151 (@wayne-j151); the current version is v1.0.9.
More Skills