Dingyi X Operator

中文说明

Cross-platform X/Twitter operator skill for Codex, Claude Code, and OpenClaw.

dingyi-x-operator is an opinionated X/Twitter operating agent for people who care about signal, taste, and reputation safety.

Best for:

• founders, builders, and operators who want help running X without turning into spam
• people who want AI help with replies, original posts, source-to-post workflows, and visual briefs
• teams or individuals who use Codex, Claude Code, or OpenClaw as their agent runtime

What it does:

• finding good tweets to engage with
• drafting replies that sound specific and human
• planning original posts and threads
• turning source material into X-ready posts
• adapting content between Chinese and English
• deciding whether a post needs a visual and writing the visual brief
• executing approved actions through opencli

Quick start

Copy the adapter for your runtime into your skills directory:

# Codex
cp -R "/path/to/repo/dingyi-x-operator/codex" "$HOME/.codex/skills/dingyi-x-operator"

# Claude Code
cp -R "/path/to/repo/dingyi-x-operator/claude-code" "$HOME/.claude/skills/dingyi-x-operator"

# OpenClaw
cp -R "/path/to/repo/dingyi-x-operator/openclaw" "/path/to/openclaw/skills/dingyi-x-operator"

Then customize local state with:

• dingyi-x-operator/templates/PROFILE.example.md
• dingyi-x-operator/templates/MEMORY.example.md

What this repository is

This repository is packaged as a single-source skill with thin adapters for:

• Codex
• Claude Code
• OpenClaw

The shared operating logic lives in:

• dingyi-x-operator/core/SKILL.md

Platform entrypoints live in:

• dingyi-x-operator/codex/SKILL.md
• dingyi-x-operator/claude-code/SKILL.md
• dingyi-x-operator/openclaw/SKILL.md

This keeps one canonical skill body in core/, while shipping installable adapter copies for each platform.

Repository layout

dingyi-x-operator/
  core/
    SKILL.md
    references/
  codex/
    SKILL.md
    agents/openai.yaml
  claude-code/
    SKILL.md
  openclaw/
    SKILL.md
  templates/
    PROFILE.example.md
    MEMORY.example.md
marketing/
  release-kit.md
  banner-brief.md

Core dependencies

Required for live X execution:

• opencli
• a running Chrome session
• the OpenCLI browser bridge extension
• a logged-in X account

Recommended helper skills when your environment supports them:

• social-media-strategist
• content-creator
• twitter-engager
• baoyu-url-to-markdown
• baoyu-translate

Installation

Recommended installation method: copy or symlink the platform adapter into your agent's skills directory.

Why this works:

• each adapter is self-contained at runtime
• symlinks are optional, not required
• the shared core/ content can regenerate adapter copies with the build script

Codex

cp -R "/path/to/repo/dingyi-x-operator/codex" "$HOME/.codex/skills/dingyi-x-operator"

Claude Code

cp -R "/path/to/repo/dingyi-x-operator/claude-code" "$HOME/.claude/skills/dingyi-x-operator"

OpenClaw

Install the openclaw/ adapter into your OpenClaw skills directory using your preferred local path convention.

Example:

cp -R "/path/to/repo/dingyi-x-operator/openclaw" "/path/to/openclaw/skills/dingyi-x-operator"

If you prefer symlinks instead of copies, that also works.

Customization

Do not commit your real operating memory or account-specific preferences.

Start from:

• dingyi-x-operator/templates/PROFILE.example.md
• dingyi-x-operator/templates/MEMORY.example.md

Typical local locations:

• $X_OPERATOR_STATE_DIR/PROFILE.md
• $X_OPERATOR_STATE_DIR/MEMORY.md
• .x-operator/PROFILE.md
• .x-operator/MEMORY.md
• ~/.x-operator/PROFILE.md
• ~/.x-operator/MEMORY.md
• .dingyi-x-operator/PROFILE.md
• .dingyi-x-operator/MEMORY.md
• ~/.dingyi-x-operator/PROFILE.md
• ~/.dingyi-x-operator/MEMORY.md

Recommended for new installs:

• use .x-operator/ or ~/.x-operator/
• use X_OPERATOR_STATE_DIR when you want per-project or per-account isolation
• keep dingyi-x-operator paths only for backward compatibility

Design principles

• credibility over volume
• shortlist before public actions
• preflight before write operations
• exact wording approval before posting
• visual planning is built in
• image generation is optional execution, not core identity

Included references

The core skill ships with reusable references for:

• execution modes
• profile template
• memory template
• output templates
• visual strategy

Launch assets

This repository also ships lightweight launch copy and visual planning files:

• marketing/release-kit.md
• marketing/banner-brief.md

Maintenance

If you update the shared skill in core/, regenerate the installable adapters with:

python3 dingyi-x-operator/scripts/build_adapters.py

License

This repository is released under the MIT License.

Contributing

Contribution guidance lives in CONTRIBUTING.md.

Not included yet

• installer scripts
• CI workflows
• publishing automation

Those can be added later once the repository shape is stable.