Tencent EdgeOne Agent Plugins
中文 | English
Tencent EdgeOne Agent Plugins is a collection of tools that help AI Agents perform configuration, operations, and troubleshooting on the Tencent Cloud EdgeOne platform.
Installation
Claude Code / Codebuddy
Install via the plugin marketplace:
# 1. Register the marketplace
/plugin marketplace add git@github.com:Tencent/tencent-edgeone-agent-plugins.git
# 2. Install the plugin
/plugin install edgeone@tencent-edgeoneOpenAI Codex / Gemini CLI / Cursor / OpenCode / Other Agents
For AI tools without a built-in plugin marketplace, tell your Agent:
Please fetch and follow the instructions in:
https://raw.githubusercontent.com/Tencent/tencent-edgeone-agent-plugins/main/INSTALL.mdYou can also clone the repository and install manually:
git clone https://github.com/Tencent/tencent-edgeone-agent-plugins.git
# Copy to the appropriate directory for your Agent tool
# e.g.: cp -r tencent-edgeone-agent-plugins/skills/tencent-edgeone-skill ~/.claude/skills/For detailed installation instructions, see INSTALL.md.
Skills
tencent-edgeone-skill
A comprehensive Skill for Tencent Cloud EdgeOne that supports site access, domain configuration, TLS certificate management, cache control, security analysis, and observability.
> This Skill uses the Tencent Cloud CLI to call Tencent Cloud APIs. The Tencent Cloud CLI must be installed before use — if it is not yet installed, the Agent will automatically install it after the Skill starts.
Included modules:
| Module | Scenarios | Typical Triggers | |---|---|---| | Acceleration | <ul><li>Site onboarding wizard</li><li>Cache purge / prefetch</li><li>Certificate automation</li></ul> | <ul><li>"Onboard example.com to EO"</li><li>"Purge all cache under /static/"</li><li>"Certificate is expiring, help me renew"</li></ul> | | Security | <ul><li>Security report generation</li><li>Security template coverage audit</li><li>Domain blocklist IP group identification</li><li>Threat IP analysis & blocking</li></ul> | <ul><li>"Generate a security status report for this week"</li><li>"Which domains have no security template?"</li><li>"Check the blocklist IP groups for this domain"</li><li>"Analyze recent attack IPs and block them"</li></ul> | | Observability | <ul><li>Traffic trend daily report</li><li>Origin health check</li><li>Offline log download</li><li>Log parsing & fault analysis</li></ul> | <ul><li>"Generate yesterday's traffic report"</li><li>"Is the origin down?"</li><li>"Download logs for example.com from yesterday afternoon"</li><li>"Too many 502s, help me analyze the logs"</li></ul> |
Skill File Structure
The edgeone skill uses a multi-level index that guides the Agent to read documents on demand:
tencent-edgeone-skill/
├── SKILL.md # tencent-edgeone-skill skill entry point
└── references/ # skill reference docs
├── api/ # API calling module
│ ├── README.md # module entry
│ ├── install.md # scenario: tccli installation
│ └── ...
├── acceleration/ # site acceleration module
│ ├── README.md # module entry
│ ├── cache-purge.md # scenario: cache purge / prefetch
│ └── ...
├── security/ # security protection module
│ ├── README.md # module entry
│ ├── security-report.md # scenario: security report
│ └── ...
└── ... # other modules...Security Design
- Write operations require user confirmation: All write operations (create, modify, delete, etc.) will explain the impact and wait for user confirmation before execution.
- Never asks for secret keys: The Agent will never ask for your SecretId / SecretKey — it uses browser-based OAuth for secure login.
- Read-only by default: Security-related Skills default to read-only queries and do not perform any modifications.
