Installable skillMIT

Twitter Research

An MCP server that provides on-demand skill loading for AI coding assistants.

BY modbenderUPDATED Oct 28, 2026
INTENDED FOR◈OpenClawInstall in a listed agent.

instructions.md

---
name: agentbox-twitter-694ff068158d
description: "Twitter/X research via paid API: search tweets with 50+ operators, fetch tweets with threads/replies/quotes, get user profiles with tweets/followers/following. Uses x402_payment tool for automatic USDC micropayments ($0.003/call). Use when: (1) searching tweets by keyword, user, or advanced operators, (2) fetching specific tweets by ID/URL with context, (3) looking up user profiles and their activity."
metadata: {"openclaw": {"emoji": "🐦", "requires": {"bins": ["openclaw"]}}}
allowed-tools: ["x402_payment"]
---

# Twitter Research

Paid Twitter/X data API at `https://twitter.x402.agentbox.fyi`. Costs $0.003 USDC per call via x402 on Solana. Use the `x402_payment` tool for all requests.

## Endpoints

### Search Tweets

Find tweets matching a query with 50+ advanced operators.

```
x402_payment({
  "url": "https://twitter.x402.agentbox.fyi/search",
  "method": "GET",
  "params": "{\"q\": \"from:elonmusk AI\", \"type\": \"Latest\", \"limit\": 20}"
})
```

**Parameters:**

| Param | Type | Default | Description |
|-------|------|---------|-------------|
| q | string | required | Search query with operators |
| type | "Latest" or "Top" | Latest | Sort by recency or popularity |
| limit | 1-200 | 20 | Max tweets to return |
| cursor | string | - | Pagination cursor from previous response |

**Search operators:**

| Operator | Example | Description |
|----------|---------|-------------|
| from:user | `from:elonmusk` | Tweets by user |
| to:user | `to:elonmusk` | Replies to user |
| @user | `@solana` | Mentioning user |
| min_faves:N | `min_faves:100` | Minimum likes |
| min_retweets:N | `min_retweets:50` | Minimum retweets |
| min_replies:N | `min_replies:10` | Minimum replies |
| filter:media | `filter:media` | Has media |
| filter:images | `filter:images` | Has images |
| filter:videos | `filter:videos` | Has video |
| filter:links | `filter:links` | Has links |
| filter:replies | `filter:replies` | Only replies |
| -filter:replies | `-filter:replies` | Exclude replies |
| since:date | `since:2025-01-01` | After date |
| until:date | `until:2025-12-31` | Before date |
| within_time:Nh | `within_time:24h` | Last N hours |
| lang:code | `lang:en` | Language filter |
| conversation_id:ID | `conversation_id:123` | In conversation |
| filter:self_threads | `from:user filter:self_threads` | User's thread replies |

Combine operators: `from:elonmusk AI min_faves:100 -filter:replies within_time:7d`

**Response:** `{ query, tweets[], count, hasMore, nextCursor, summary }`

The `summary` object includes: `totalLikes`, `totalRetweets`, `totalReplies`, `avgLikes`, `maxLikes`, `topTweetId`.

### Fetch Tweets

Get tweet(s) by ID or URL with optional thread, replies, and quotes.

```
x402_payment({
  "url": "https://twitter.x402.agentbox.fyi/tweet/1585841080431321088",
  "method": "GET",
  "params": "{\"include\": \"thread,replies\", \"limit\": 50}"
})
```

**URL format:** `/tweet/{ref}` where ref is:
- Tweet ID: `1585841080431321088`
- Comma-separated IDs: `123,456,789` (batch fetch, no includes)

**Parameters:**

| Param | Type | Default | Description |
|-------|------|---------|-------------|
| include | string | - | Comma-separated: `thread`, `replies`, `quotes` |
| limit | 1-200 | 50 | Max replies/quotes to return |

**Include options (single tweet only):**
- `thread` - author's self-reply chain in the conversation
- `replies` - replies from other users
- `quotes` - quote tweets of this tweet

**Response:** `{ tweets[], count, parent?, thread?, replies?, quotes? }`

If the fetched tweet is a reply, `parent` is auto-included with the replied-to tweet.

### Fetch Users

Get user profile(s) with optional tweets, followers, or following.

```
x402_payment({
  "url": "https://twitter.x402.agentbox.fyi/user/elonmusk",
  "method": "GET",
  "params": "{\"include\": \"tweets\", \"limit\": 20}"
})
```

**URL format:** `/user/{ref}` where ref is:
- Username: `elonmusk` or `@elonmusk`
- User ID: `44196397`
- Comma-separated IDs: `123,456` (batch fetch, no includes)

**Parameters:**

| Param | Type | Default | Description |
|-------|------|---------|-------------|
| include | string | - | Comma-separated: `tweets`, `followers`, `following` |
| limit | 1-200 | 50 | Max tweets/followers/following to return |
| include_replies | boolean | false | Include replies in user's tweets |

**Response (single):** `{ user, tweets?, followers?, following? }`
**Response (batch):** `{ users[], count, hasMore }`

## Usage Patterns

### Monitor a topic

```
x402_payment({
  "url": "https://twitter.x402.agentbox.fyi/search",
  "method": "GET",
  "params": "{\"q\": \"x402 OR \\\"402 payment\\\" min_faves:5 within_time:24h\", \"type\": \"Latest\"}"
})
```

### Get a tweet with full context

```
x402_payment({
  "url": "https://twitter.x402.agentbox.fyi/tweet/1585841080431321088",
  "method": "GET",
  "params": "{\"include\": \"thread,replies,quotes\", \"limit\": 20}"
})
```

### Research a user

```
x402_payment({
  "url": "https://twitter.x402.agentbox.fyi/user/CoinbaseDev",
  "method": "GET",
  "params": "{\"include\": \"tweets\", \"limit\": 50}"
})
```

### Paginate results

Use `nextCursor` from a previous response:

```
x402_payment({
  "url": "https://twitter.x402.agentbox.fyi/search",
  "method": "GET",
  "params": "{\"q\": \"from:elonmusk\", \"cursor\": \"DAABCgAB...\"}"
})
```

## Cost

Every call costs $0.003 USDC on Solana mainnet. Each paginated request is a separate call. Plan queries to minimize calls - use specific operators and reasonable limits.

## Errors

| HTTP | Meaning |
|------|---------|
| 400 | Invalid parameters (check query syntax) |
| 402 | Payment required (handled automatically by x402_payment) |
| 404 | Tweet or user not found |
| 502 | Upstream API error |
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.