agentleFS
mcarmona1064-cellmcarmona1064-cell/clawbridge-codex/.claude/skills/x-integration/SKILL.md

x-integration

Skill1 starsChanged 6 months agoMITView on GitHub

X (Twitter) integration for ClawBridge. Post tweets, like, reply, retweet, and quote. Use for setup, testing, or troubleshooting X functionality. Triggers on "setup x", "x integration", "twitter", "post tweet", "tweet".

  • Reads credentials
  • Installs packages

8 other projects use a version of this, 8 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use add-gmail, add-telegram-swarm and 1 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • jaypylesSkill

    add-gmail

    Add Gmail integration to NanoClaw. Can be configured as a tool (agent reads/sends emails when triggered from Discord) or as a full channel (emails can trigger the agent, schedule tasks, and receive replies). Guides through GCP OAuth setup and implements the integration.

    Also in 5 projects that use this

  • juancruzmunozalbeloSkill

    add-telegram-swarm

    Add Agent Swarm (Teams) support to Telegram. Each subagent gets its own bot identity in the group. Requires Telegram channel to be set up first (use /add-telegram). Triggers on "agent swarm", "agent teams telegram", "telegram swarm", "bot pool".

    Also in 5 projects that use this

  • juancruzmunozalbeloSkill

    add-voice-transcription

    Add voice message transcription to NanoClaw using OpenAI's Whisper API. Automatically transcribes WhatsApp voice notes so the agent can read and respond to them.

    Also in 4 projects that use this

The file

We split this skill into the rules it gives an agent

11 rules. 6 are common, 0 are distinctive, and 5 are very specific to this project.

  • 6

    Common

    Most of the 8 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Common:

    Before using this skill, ensure:

    line 24
  2. Common:

    Add import after other local imports:

    line 163
  3. Very specific:

    Add import after cron-parser import:

    line 186
  4. Very specific:

    Add to the end of tools array (before the closing ]):

    line 192
  5. Very specific:

    All paths below are relative to project root (CLAWBRIDGE_ROOT).

    line 236
  6. Common:

    Note: Only the main group can use X tools.

    line 301
  7. Common:

    If Chrome fails to launch:

    line 354
  8. Very specific:

    Default timeout is 2 minutes (120s).

    line 374
  9. Common:

    If X updates their UI, selectors in scripts may break.

    line 385
  10. Common:

    If MCP tools not found in container:

    line 402
  11. Very specific:

    Only main group can use X tools (enforced in agent.ts and host.ts)

    line 416

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

X (Twitter) Integration

Browser automation for X interactions via WhatsApp.

Compatibility: ClawBridge v1.0.0. Directory structure may change in future versions.

Features

ActionToolDescription
Postx_postPublish new tweets
Likex_likeLike any tweet
Replyx_replyReply to tweets
Retweetx_retweetRetweet without comment
Quotex_quoteQuote tweet with comment

Prerequisites

Before using this skill, ensure:

  1. ClawBridge is installed and running - WhatsApp connected, service active
  2. Dependencies installed:
    pnpm ls playwright dotenv-cli || pnpm install playwright dotenv-cli
  3. CHROME_PATH configured in .env (if Chrome is not at default location):
    # Find your Chrome path
    mdfind "kMDItemCFBundleIdentifier == 'com.google.Chrome'" 2>/dev/null | head -1
    # Add to .env
    CHROME_PATH=/path/to/Google Chrome.app/Contents/MacOS/Google Chrome

Quick Start

# 1. Setup authentication (interactive)
pnpm exec dotenv -e .env -- pnpm exec tsx .claude/skills/x-integration/scripts/setup.ts
# Verify: data/x-auth.json should exist after successful login

# 2. Rebuild container to include skill
./container/build.sh
# Verify: Output shows "COPY .claude/skills/x-integration/agent.ts"

# 3. Rebuild host and restart service
pnpm run build
launchctl kickstart -k gui/$(id -u)/com.clawbridge  # macOS
# Linux: systemctl --user restart clawbridge
# Verify: launchctl list | grep clawbridge (macOS) or systemctl --user status clawbridge (Linux)

Configuration

Environment Variables

VariableDefaultDescription
CHROME_PATH/Applications/Google Chrome.app/Contents/MacOS/Google ChromeChrome executable path
CLAWBRIDGE_ROOTprocess.cwd()Project root directory
LOG_LEVELinfoLogging level (debug, info, warn, error)

Set in .env file (loaded via dotenv-cli at runtime):

# .env
CHROME_PATH=/Applications/Google Chrome.app/Contents/MacOS/Google Chrome

Configuration File

Edit lib/config.ts to modify defaults:

export const config = {
    // Browser viewport
    viewport: { width: 1280, height: 800 },

    // Timeouts (milliseconds)
    timeouts: {
        navigation: 30000,    // Page navigation
        elementWait: 5000,    // Wait for element
        afterClick: 1000,     // Delay after click
        afterFill: 1000,      // Delay after form fill
        afterSubmit: 3000,    // Delay after submit
        pageLoad: 3000,       // Initial page load
    },

    // Tweet limits
    limits: {
        tweetMaxLength: 280,
    },
};

Data Directories

Paths relative to project root:

PathPurposeGit
data/x-browser-profile/Chrome profile with X sessionIgnored
data/x-auth.jsonAuth state markerIgnored
logs/clawbridge.logService logs (contains X operation logs)Ignored

Architecture

┌─────────────────────────────────────────────────────────────┐
│  Container (Linux VM)                                       │
│  └── agent.ts → MCP tool definitions (x_post, etc.)    │
│      └── Writes IPC request to /workspace/ipc/tasks/       │
└──────────────────────┬──────────────────────────────────────┘
                       │ IPC (file system)
                       ▼
┌─────────────────────────────────────────────────────────────┐
│  Host (macOS)                                               │
│  └── src/ipc.ts → processTaskIpc()                         │
│      └── host.ts → handleXIpc()                         │
│          └── spawn subprocess → scripts/*.ts               │
│              └── Playwright → Chrome → X Website           │
└─────────────────────────────────────────────────────────────┘

Why This Design?

  • API is expensive - X official API requires paid subscription ($100+/month) for posting
  • Bot browsers get blocked - X detects and bans headless browsers and common automation fingerprints
  • Must use user's real browser - Reuses the user's actual Chrome on Host with real browser fingerprint to avoid detection
  • One-time authorization - User logs in manually once, session persists in Chrome profile for future use

File Structure

.claude/skills/x-integration/
├── SKILL.md          # This documentation
├── host.ts           # Host-side IPC handler
├── agent.ts          # Container-side MCP tool definitions
├── lib/
│   ├── config.ts     # Centralized configuration
│   └── browser.ts    # Playwright utilities
└── scripts/
    ├── setup.ts      # Interactive login
    ├── post.ts       # Post tweet
    ├── like.ts       # Like tweet
    ├── reply.ts      # Reply to tweet
    ├── retweet.ts    # Retweet
    └── quote.ts      # Quote tweet

Integration Points

To integrate this skill into ClawBridge, make the following modifications:


1. Host side: src/ipc.ts

Add import after other local imports:

import { handleXIpc } from '../.claude/skills/x-integration/host.js';

Modify processTaskIpc function's switch statement default case:

// Find:
default:
logger.warn({ type: data.type }, 'Unknown IPC task type');

// Replace with:
default:
const handled = await handleXIpc(data, sourceGroup, isMain, DATA_DIR);
if (!handled) {
    logger.warn({ type: data.type }, 'Unknown IPC task type');
}

2. Container side: container/agent-runner/src/ipc-mcp.ts

Add import after cron-parser import:

// @ts-ignore - Copied during Docker build from .claude/skills/x-integration/
import { createXTools } from './skills/x-integration/agent.js';

Add to the end of tools array (before the closing ]):

    ...createXTools({ groupFolder, isMain })

3. Build script: container/build.sh

Change build context from container/ to project root (required to access .claude/skills/):

# Find:
docker build -t "${IMAGE_NAME}:${TAG}" .

# Replace with:
cd "$SCRIPT_DIR/.."
docker build -t "${IMAGE_NAME}:${TAG}" -f container/Dockerfile .

4. Dockerfile: container/Dockerfile

First, update the build context paths (required to access .claude/skills/ from project root):

# Find:
COPY agent-runner/package*.json ./
...
COPY agent-runner/ ./

# Replace with:
COPY container/agent-runner/package*.json ./
...
COPY container/agent-runner/ ./

Then add COPY line after COPY container/agent-runner/ ./ and before RUN pnpm run build:

# Copy skill MCP tools
COPY .claude/skills/x-integration/agent.ts ./src/skills/x-integration/

Setup

All paths below are relative to project root (CLAWBRIDGE_ROOT).

1. Check Chrome Path

# Check if Chrome exists at configured path
cat .env | grep CHROME_PATH
ls -la "$(grep CHROME_PATH .env | cut -d= -f2)" 2>/dev/null || \
echo "Chrome not found - update CHROME_PATH in .env"

2. Run Authentication

pnpm exec dotenv -e .env -- pnpm exec tsx .claude/skills/x-integration/scripts/setup.ts

This opens Chrome for manual X login. Session saved to data/x-browser-profile/.

Verify success:

cat data/x-auth.json  # Should show {"authenticated": true, ...}

3. Rebuild Container

./container/build.sh

Verify success:

./container/build.sh 2>&1 | grep -i "agent.ts"  # Should show COPY line

4. Restart Service

pnpm run build
launchctl kickstart -k gui/$(id -u)/com.clawbridge  # macOS
# Linux: systemctl --user restart clawbridge

Verify success:

launchctl list | grep clawbridge  # macOS — should show PID and exit code 0 or -
# Linux: systemctl --user status clawbridge

Usage via WhatsApp

Replace @Assistant with your configured trigger name (ASSISTANT_NAME in .env):

@Assistant post a tweet: Hello world!

@Assistant like this tweet https://x.com/user/status/123

@Assistant reply to https://x.com/user/status/123 with: Great post!

@Assistant retweet https://x.com/user/status/123

@Assistant quote https://x.com/user/status/123 with comment: Interesting

Note: Only the main group can use X tools. Other groups will receive an error.

Testing

Scripts require environment variables from .env. Use dotenv-cli to load them:

Check Authentication Status

# Check if auth file exists and is valid
cat data/x-auth.json 2>/dev/null && echo "Auth configured" || echo "Auth not configured"

# Check if browser profile exists
ls -la data/x-browser-profile/ 2>/dev/null | head -5

Re-authenticate (if expired)

pnpm exec dotenv -e .env -- pnpm exec tsx .claude/skills/x-integration/scripts/setup.ts

Test Post (will actually post)

echo '{"content":"Test tweet - please ignore"}' | pnpm exec dotenv -e .env -- pnpm exec tsx .claude/skills/x-integration/scripts/post.ts

Test Like

echo '{"tweetUrl":"https://x.com/user/status/123"}' | pnpm exec dotenv -e .env -- pnpm exec tsx .claude/skills/x-integration/scripts/like.ts

Or export CHROME_PATH manually before running:

export CHROME_PATH="/path/to/chrome"
echo '{"content":"Test"}' | pnpm exec tsx .claude/skills/x-integration/scripts/post.ts

Troubleshooting

Authentication Expired

pnpm exec dotenv -e .env -- pnpm exec tsx .claude/skills/x-integration/scripts/setup.ts
launchctl kickstart -k gui/$(id -u)/com.clawbridge  # macOS
# Linux: systemctl --user restart clawbridge

Browser Lock Files

If Chrome fails to launch:

rm -f data/x-browser-profile/SingletonLock
rm -f data/x-browser-profile/SingletonSocket
rm -f data/x-browser-profile/SingletonCookie

Check Logs

# Host logs (relative to project root)
grep -i "x_post\|x_like\|x_reply\|handleXIpc" logs/clawbridge.log | tail -20

# Script errors
grep -i "error\|failed" logs/clawbridge.log | tail -20

Script Timeout

Default timeout is 2 minutes (120s). Increase in host.ts:

const timer = setTimeout(() => {
  proc.kill('SIGTERM');
  resolve({ success: false, message: 'Script timed out (120s)' });
}, 120000);  // ← Increase this value

X UI Selector Changes

If X updates their UI, selectors in scripts may break. Current selectors:

ElementSelector
Tweet input[data-testid="tweetTextarea_0"]
Post button[data-testid="tweetButtonInline"]
Reply button[data-testid="reply"]
Like[data-testid="like"]
Unlike[data-testid="unlike"]
Retweet[data-testid="retweet"]
Unretweet[data-testid="unretweet"]
Confirm retweet[data-testid="retweetConfirm"]
Modal dialog[role="dialog"][aria-modal="true"]
Modal submit[data-testid="tweetButton"]

Container Build Issues

If MCP tools not found in container:

# Verify build copies skill
./container/build.sh 2>&1 | grep -i skill

# Check container has the file
docker run clawbridge-agent ls -la /app/src/skills/

Security

  • data/x-browser-profile/ - Contains X session cookies (in .gitignore)
  • data/x-auth.json - Auth state marker (in .gitignore)
  • Only main group can use X tools (enforced in agent.ts and host.ts)
  • Scripts run as subprocesses with limited environment

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.