agent-buzz

<!-- autoresearch: variation B — sharper output via clustering, signal score, insight extraction, and skip-gates --> > **${var}** — Specific project or topic to prioritize (e.g. "MCP protocol", "browser-use"). If empty, searches AI agents broadly. Read `memory/MEMORY.md` for context. Read the last 3 days of `memory/logs/` — extract every `https://x.com/.../status/<id>` URL already posted by this skill and treat those IDs as a dedup set. ## Goal Publish a curated, narrative-aware read on what the AI-agent scene on X talked about in the last 24h. **Curation, not aggregation.** Better to ship 6 high-signal tweets in 2 clusters than 10 tweets of mixed noise. ## Steps ### 1. Fetch candidates ```bash FROM_DATE=$(date -u -d "1 day ago" +%Y-%m-%d 2>/dev/null || date -u -v-1d +%Y-%m-%d) TO_DATE=$(date -u +%Y-%m-%d) ``` Issue one primary x_search call. The response for each tweet **must include** explicit engagement counts (likes, retweets, replies) and follower count if visible — without these numbers the signal scoring in step 3 cannot run. ```bash curl -s -X POST "https://api.x.ai/v1/responses" \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $XAI_API_KEY" \ -d '{ "model": "grok-4-1-fast", "input": [{"role": "user", "content": "Search X from '"$FROM_DATE"' to '"$TO_DATE"' for tweets in the AI-agents conversation: autonomous agents, agent frameworks, MCP / agent protocols, agent products, agent benchmarks, agent research papers. Return up to 40...