res-x（分辨率 x）

X/Twitter Fetch & Search

Fetch tweet content by URL and search X posts using xAI Responses API with Grok's x_search tool. Solves the problem of X/Twitter blocking WebFetch with "JavaScript disabled" errors.

Architecture

Fetching batches URLs in groups of 3 per API call to reduce cost.

Prerequisites

Required Tools

API Key

This skill requires an xAI API key. There is no fallback mode.

Keychain Setup (One-Time)

# 1. Create a dedicated keychain (skip if already exists)
security create-keychain -p 'YourPassword' ~/Library/Keychains/claude-keys.keychain-db

# 2. Add keychain to search list
security list-keychains -s ~/Library/Keychains/claude-keys.keychain-db ~/Library/Keychains/login.keychain-db /Library/Keychains/System.keychain

# 3. Store your xAI API key
echo -n "Enter xAI API key: " && read -s key && security add-generic-password -s "xai-api" -a "$USER" -w "$key" ~/Library/Keychains/claude-keys.keychain-db && unset key && echo

Before using: security unlock-keychain ~/Library/Keychains/claude-keys.keychain-db

Script Usage

# Fetch single tweet
uv run scripts/x_fetch.py fetch "https://x.com/user/status/123456"

# Fetch multiple tweets (batched, 3 per API call)
uv run scripts/x_fetch.py fetch "url1" "url2" "url3" "url4" "url5"

# Force one-per-URL for max fidelity
uv run scripts/x_fetch.py fetch "url1" "url2" --single

# Search X
uv run scripts/x_fetch.py search "query terms"
uv run scripts/x_fetch.py search "query terms" --quick

# JSON output (both commands)
uv run scripts/x_fetch.py fetch "url" --json
uv run scripts/x_fetch.py search "query" --json

Workflow

Step 0: Detect xAI Key (Mandatory)

Run before every invocation:

security find-generic-password -s "xai-api" -w ~/Library/Keychains/claude-keys.keychain-db 2>/dev/null && echo "XAI_AVAILABLE=true" || echo "XAI_AVAILABLE=false"

If XAI_AVAILABLE=false, report that this skill requires an xAI key and show the keychain setup instructions above.

Step 1: Detect Intent

• URLs present in user input or referenced note -> fetch
• Query text only -> search

Step 2: Execute

For fetch:

1. Extract all X/Twitter URLs from user input or referenced file
2. Run the script with all URLs as arguments
3. The script batches them (3 per API call) automatically

For search:

1. Run the script with the search query
2. Use --quick for fast overview, omit for deeper results

Step 3: Present Results

• For fetch: present tweet-by-tweet with full content, engagement, thread/quote context
• For search: present as a list with engagement metrics

URL Patterns Accepted

https://x.com/{user}/status/{id}
https://twitter.com/{user}/status/{id}
https://x.com/{user}/status/{id}?s=20
https://x.com/{user}/status/{id}?t=...&s=...
https://x.com/i/article/{id}

Tweets normalized to https://x.com/{user}/status/{id}, articles to https://x.com/i/article/{id} before processing.

Batch Processing

For processing saved X links from a file (e.g., Obsidian daily note):

1. Read the file content
2. Extract all X/Twitter URLs (tweets: https?://(?:x\.com|twitter\.com)/\w+/status/\d+, articles: https?://(?:x\.com|twitter\.com)/i/article/[\w\-]+)
3. Pass all URLs to the script: uv run scripts/x_fetch.py fetch "url1" "url2"...
4. Present results organized by URL

Cost

Constraints

DO:

• Run Step 0 before every invocation
• Validate URLs before calling the script
• Use --json when parsing results programmatically
• Present full tweet content without truncation

DON'T:

• Try WebFetch on X URLs (fails with JS disabled)
• Skip the xAI key check
• Use this for general web search (use res-web or res-deep)

Troubleshooting

xAI key not found:

security find-generic-password -s "xai-api" ~/Library/Keychains/claude-keys.keychain-db

If not found, run keychain setup above.

Keychain locked:

security unlock-keychain ~/Library/Keychains/claude-keys.keychain-db

Script errors: Ensure uv is installed: which uv