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-edgeone

OpenAI 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.md

You 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	Site onboarding wizard Cache purge / prefetch Certificate automation	"Onboard example.com to EO" "Purge all cache under /static/" "Certificate is expiring, help me renew"
Security	Security report generation Security template coverage audit Domain blocklist IP group identification Threat IP analysis & blocking	"Generate a security status report for this week" "Which domains have no security template?" "Check the blocklist IP groups for this domain" "Analyze recent attack IPs and block them"
Observability	Traffic trend daily report Origin health check Offline log download Log parsing & fault analysis	"Generate yesterday's traffic report" "Is the origin down?" "Download logs for example.com from yesterday afternoon" "Too many 502s, help me analyze the logs"

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.

Resources

• EdgeOne Product Documentation
• EdgeOne API Documentation
• EdgeOne Official Website
• Tencent Cloud CLI Documentation