clawbird（clawbird 搜索）

Clawbird — X/Twitter Tools

You have access to 13 tools for interacting with X (Twitter) via the official X API v2. All tools return JSON with results and estimated API cost.

Authentication & Credentials

This plugin authenticates to the X API using OAuth 1.0a User Context for write operations (posting, liking, following, DMs) and optionally a Bearer Token for read-only operations (search, user lookup).

Where credentials come from: You must generate them at the X Developer Portal:

1. Create a Project and App at developer.x.com
2. Generate OAuth 1.0a keys: API Key, API Secret, Access Token, Access Token Secret
3. Optionally generate a Bearer Token for read-only operations

How credentials are provided: Credentials are passed to the plugin at runtime by the OpenClaw plugin config system (pluginConfig). The plugin never reads config files directly. Fallback: environment variables X_API_KEY, X_API_SECRET, X_ACCESS_TOKEN, X_ACCESS_SECRET, X_BEARER_TOKEN.

No credentials are hardcoded or bundled. The plugin will return a clear error if credentials are missing.

External Endpoints

All network requests go exclusively to the official X API v2. No other hosts are contacted.

Security & Privacy

• Network access: Only api.x.com (official X API). No other domains are contacted.
• Local file access: Writes a session-scoped clawbird-interactions.jsonl file to the working directory, logging mutation actions (posts, likes, follows, DMs) so the agent can avoid duplicating work. No other files are read or written beyond the declared environment variables.
• Credential handling: OAuth tokens are read from plugin config or env vars at runtime and passed to the X API via signed HTTP headers. They are never logged, cached to disk, or transmitted to any third party.
• Data sent to X: Only the data you explicitly provide in tool parameters (tweet text, search queries, usernames, message text). No additional user data is collected or sent.
• Data received from X: Tweet content, user profiles, DM messages, and engagement metrics as returned by the X API. This data is returned to the agent as JSON and not stored.

Trust Statement

Clawbird is an npm-distributed skill — the runtime code is installed via npm i -g @xonder/clawbird and is not bundled in the ClawHub skill archive. The security properties below can be verified by auditing the source:

• Source: https://github.com/xonder/clawbird (MIT)
• npm: https://www.npmjs.com/package/@xonder/clawbird
• Makes no network requests other than to api.x.com (source: src/client.ts)
• Reads and writes one local file (clawbird-interactions.jsonl) for session interaction logging — no other filesystem access (source: src/interaction-log.ts)
• Has zero transitive dependencies beyond the official @xdevplatform/xdk SDK and @sinclair/typebox
• Includes a comprehensive test suite (200+ tests) verifiable via npm test

Since this skill installs code from npm at runtime, review the source or pin a specific version (npm i -g @xonder/clawbird@1.1.0) before granting credentials.

Write Actions & Autonomous Use

The following tools modify remote state on your X account:

Recommendation: If running autonomously, consider requiring explicit user confirmation before write actions by configuring agent-level tool policies. Read-only tools (x_get_tweet, x_search_tweets, x_get_user_profile, x_get_mentions, x_get_dms, x_get_cost_summary) are safe for autonomous use.

Available Tools

Posting

x_post_tweet — Post a single tweet.

• text (required): Tweet content (max 280 chars)
• Returns: { id, text, url, estimatedCost }

x_post_thread — Post a multi-tweet thread.

• tweets (required): Array of tweet texts (posted in order, each as a reply to the previous)
• Returns: { threadId, tweetCount, tweets: [{ id, text, url }], estimatedCost }

x_reply_tweet — Reply to an existing tweet.

• tweetId (required): Tweet ID or full URL (e.g. https://x.com/user/status/123456)
• text (required): Reply content (max 280 chars)
• Returns: { id, text, url, inReplyTo, estimatedCost }

Engagement

x_like_tweet — Like a tweet.

• tweetId (required): Tweet ID or full URL
• Returns: { liked, tweetId, estimatedCost }

Social

x_follow_user — Follow a user.

• username (required): Username to follow (with or without @)
• Returns: { following, user: { id, username }, estimatedCost }

Research

x_get_tweet — Get a single tweet by ID or URL.

• tweetId (required): Tweet ID or full URL (e.g. https://x.com/user/status/123456)
• Returns: { id, text, authorId, createdAt, metrics, conversationId, lang, url, author: { id, name, username, verified, profileImageUrl }, estimatedCost }

x_search_tweets — Search recent tweets (last 7 days).

• query (required): Search query — supports X operators like from:user, #hashtag, "exact phrase", -exclude, lang:en
• maxResults (optional): 10–100, default 10
• Returns: { query, resultCount, tweets: [{ id, text, authorId, createdAt, metrics, url }], estimatedCost }

x_get_user_profile — Get a user's profile.

• username (required): Username with or without @
• Returns: { id, name, username, description, followersCount, followingCount, tweetCount, verified, profileImageUrl, url, createdAt, location, profileUrl, estimatedCost }

x_get_mentions — Get recent mentions of the authenticated account.

• maxResults (optional): 5–100, default 10
• Returns: { resultCount, mentions: [{ id, text, authorId, createdAt, metrics, url }], estimatedCost }

Direct Messages

x_send_dm — Send a direct message to a user.

• username (required): Recipient's username (with or without @)
• text (required): Message content
• Returns: { sent, eventId, conversationId, recipient: { id, username }, estimatedCost }

x_get_dms — Get recent direct messages.

• username (optional): Filter DMs to a specific user's conversation
• maxResults (optional): 1–100, default 10
• Returns: { resultCount, messages: [{ id, text, senderId, createdAt, conversationId, eventType }], estimatedCost }

Utility

x_get_interaction_log — Get the log of all write actions performed this session (posts, replies, likes, follows, DMs). Useful to review what has already been done and avoid duplicating actions.

• limit (optional): Maximum number of recent entries to return (default: all)
• Returns: { totalEntries, returned, logFile, entries: [{ timestamp, action, summary, details }] }

x_get_cost_summary — Get cumulative API cost for this session.

• No parameters required
• Returns: { totalCost, breakdown: { [action]: { calls, totalCost } } }

Best Practices

Search Queries

• Use from:username to search a specific user's tweets
• Use #hashtag for hashtag search
• Use "exact phrase" for exact matches
• Combine operators: #AI from:openai -is:retweet lang:en
• Use -is:retweet to filter out retweets

Thread Formatting

• Keep each tweet under 280 characters
• Start with a strong hook in tweet 1
• Number tweets (1/N) for long threads
• End with a call to action or summary

Cost Awareness

Every tool response includes an estimatedCost field. Approximate costs:

• Post/Reply: ~$0.01 per tweet
• Like: ~$0.005
• Search: ~$0.005 per result
• User lookup: ~$0.001
• Mentions: ~$0.005 per result
• Send DM: ~$0.01
• Read DMs: ~$0.005 per result
• Get tweet: ~$0.005

Use x_get_cost_summary to check cumulative session spend before expensive operations.

Rate Limits

• Posting: 200 tweets per 15 minutes
• Search: 180 requests per 15 minutes (user), 450 (app)
• Likes: 50 per 15 minutes
• User lookup: 900 per 15 minutes
• Mentions: 180 per 15 minutes
• DMs: 200 messages per 15 minutes, 1000 per 24 hours
• Following: 400 per 24 hours

Error Handling

All tools return errors as { error: "message", details?: ... }. Common issues:

• Rate limiting (wait and retry)
• Authentication errors (check API credentials)
• Tweet not found (verify ID/URL)
• Empty text (provide non-empty content)