fetch-tweets

<!-- autoresearch: variation B — sharper output via clustering + signal line + insight extraction --> > **${var}** — Search query for X/Twitter. **Required** — set your query in aeon.yml. Today is ${today}. Search X for tweets matching **${var}** and produce a *curated* digest grouped by sub-narrative — not a flat chronological list. ## Steps 1. **Load previously-reported tweet URLs** from two sources, then union them into `SEEN_TWEETS`: - **Persistent seen-file** (`memory/fetch-tweets-seen.txt`) — if it exists, read all URLs. - **Last 3 days of `memory/logs/`** — grep each log file for lines matching `https://x.com/`. You'll use `SEEN_TWEETS` in step 5 to filter duplicates. 2. **Build the search prompt for Grok.** Pass `${var}` to Grok **verbatim** as the search query — let Grok interpret OR/AND operators in the var as-is. Do NOT narrow it to a single angle; broad coverage is the goal. Ask Grok to return **at least 15–20 candidate tweets** (you'll cull to ~7–10 in curation). Always require explicit engagement counts (likes, retweets, replies) so ranking is data-driven, not vibes. 3. **Search tweets.** Use whichever path is available, and **record which path produced the result** (you'll log it in step 6 as `SOURCE_PATH=cache|api|websearch`): **Path A — pre-fetched cache** (preferred, when the workflow ran `scripts/prefetch-xai.sh`): ```bash cat .xai-cache/fetch-tweets.json 2>/dev/null | jq -r '.output[] | select(.type == "message") | .content[] |...