agent-buzz

name	Agent Buzz
description	Top 10 tweets by influence mentioning AI agents or OpenClaw
tags	["social"]

Read memory/MEMORY.md for context. Read the last 2 days of memory/logs/ to avoid repeating tweets.

Steps

1. Search X for high-engagement tweets about AI agents and OpenClaw using the X.AI API:

   FROM_DATE=$(date -u -d "2 days ago" +%Y-%m-%d 2>/dev/null || date -u -v-2d +%Y-%m-%d)
   TO_DATE=$(date -u +%Y-%m-%d)
   
   # Search AI agents broadly
   curl -s -X POST "https://api.x.ai/v1/responses" \
     -H "Content-Type: application/json" \
     -H "Authorization: Bearer $XAI_API_KEY" \
     -d '{
       "model": "grok-4-1-fast",
       "input": [{"role": "user", "content": "Search X for the most viral and influential tweets from '"$FROM_DATE"' to '"$TO_DATE"' mentioning AI agents, autonomous agents, agent frameworks, agent protocols, or agentic AI. Rank by influence — prioritize accounts with large followings, high engagement (likes, retweets, quote tweets), and people who are builders/founders/researchers in the space. Return the top 15 tweets. For each: @handle, follower count if visible, a one-line summary of what they said, engagement stats if available, and the direct link (https://x.com/username/status/ID)."}],
       "tools": [{"type": "x_search", "from_date": "'"$FROM_DATE"'", "to_date": "'"$TO_DATE"'"}]
     }'
   
   # Search OpenClaw specifically
   curl -s -X POST "https://api.x.ai/v1/responses" \
     -H "Content-Type: application/json" \
     -H "Authorization: Bearer $XAI_API_KEY" \
     -d '{
       "model": "grok-4-1-fast",
       "input": [{"role": "user", "content": "Search X for all tweets from '"$FROM_DATE"' to '"$TO_DATE"' mentioning OpenClaw, OpenClaw-RL, or open-claw. Return every tweet you find with @handle, summary, engagement, and direct link (https://x.com/username/status/ID)."}],
       "tools": [{"type": "x_search", "from_date": "'"$FROM_DATE"'", "to_date": "'"$TO_DATE"'"}]
     }'
   

   If XAI_API_KEY is not set or the curl command fails (sandbox blocking, network errors), fall back to WebSearch. IMPORTANT: Always include "today" or "past 24 hours" and the current date in your search query to force fresh results — e.g. "AI agents twitter today ${today}", "OpenClaw twitter ${today}". Discard any results older than 48 hours. Summarize what you find.

2. Merge both result sets. Deduplicate against recent logs. Rank by influence:

   • Follower count / account credibility (builders > commentators > anons)
   • Engagement (likes, RTs, quote tweets)
   • Substance (actual insight > hype > shilling)
   • OpenClaw mentions get priority — always include them in the top 10 if any exist

3. Pick the top 10 tweets.

4. Send via ./notify (under 4000 chars):

   *Agent Buzz — ${today}*
   
   1. @handle — one-line summary
      [link](https://x.com/...)
   
   2. @handle — one-line summary
      [link](https://x.com/...)
   
   ...
   
   *OpenClaw mentions:* N found (or "none today")
   

5. Log to memory/logs/${today}.md. If nothing found, log "AGENT_BUZZ_OK" and end.

Environment Variables Required

• XAI_API_KEY — X.AI API key for Grok x_search