curation

Selection and ranking rules — how to turn the raw fetched pool into the handful of items the reader actually wants, guided by the taste profile in…

npx skills add https://github.com/google-gemini/gemini-managed-agents-templates --skill curation

Curation Skill

Curation is editing, not summarizing. From the candidate pool in new_items.json, choose the items this specific reader should see today.

Selection Rules

  1. The taste profile decides. Re-read interests.json before selecting. preferences.more entries boost an item; preferences.less entries drop it unless it is genuinely major; preferences.notes carry dated, specific guidance from past feedback — treat recent notes as the strongest signal.
  2. Cap the briefing at 10 news items total across all topics, and up to 4 community-buzz items. Fewer is fine; padding is not. If a topic produced nothing worth including, say so in one line rather than including filler.
  3. Prefer, in order: primary sources and official announcements → original reporting → high-engagement community discussion → aggregation/commentary. Drop listicles and rumor pieces by default (they are on the default less list).
  4. Collapse duplicates of one story. When several outlets carry the same development, pick the strongest single link and mention coverage breadth in the why-it-matters line ("also covered by Reuters and The Verge") — breadth itself is a signal the story matters.
  5. Genre feeds are the serendipity channel. Pick at most 2-3 genre items the reader did not explicitly ask about but would plausibly care about, given their profile. Label them as discoveries.
  6. Community items are for signal, not noise. Select a Hacker News item only when it adds something the news coverage lacks: practitioner experience, technical pushback, or early buzz on something not yet covered.
  7. Longform earns its slot with depth. Select at most 2 Medium pieces per briefing, and only for genuine substance — tutorials, architecture write-ups, first-hand experience reports. Medium has no engagement signal in the feed, so judge entirely on title substance and taste-profile fit; skip anything that reads as engagement-bait or paywalled teaser fluff.

Ranking Within the Briefing

Lead with the item that best combines: relevance to an explicit topic, recency, source quality, and taste-profile fit. Genre discoveries and community buzz always come after the reader's explicit topics.

What Never Gets Selected

  • Items you cannot link.
  • Items whose headline alone is too thin to say why it matters — do not pad with invented detail.
  • Pure engagement-bait framed as news, regardless of topic match.

More skills from google-gemini

agent-tui
google-gemini
Main Agents: Do NOT use this skill directly. If you need to test the TUI, invoke the `tui_tester` subagent. Drive terminal UI (TUI) applications…
gemini-api-cli
google-gemini
Guide for using the Gemini API CLI tool. Use when you need to interact with the Gemini API via the command line, manage agents, or generate media (images,…
behavioral-evals
google-gemini
Guidance for creating, running, fixing, and promoting behavioral evaluations. Use when verifying agent decision logic, debugging failures, debugging prompt…
gemini-live-api-dev
google-gemini
Real-time bidirectional streaming with Gemini over WebSockets for audio, video, and text conversations. Supports audio input/output (16 kHz PCM), video frames, text, and automatic transcriptions with voice activity detection for interruption handling Includes native audio features: affective dialog, proactive audio, and thinking mode; function calling for synchronous and asynchronous tool use; and Google Search grounding Offers session management with context compression, resumption, and...
gemini-omni-flash-api
google-gemini
Use this skill for generative video editing, text-to-video, image-referenced video generation, and first-frame-to-video transition animations using the…
gemini-api-dev
google-gemini
Build applications with Google's Gemini models, supporting multimodal content, function calling, and structured outputs across Python, JavaScript, Go, and Java. Access current Gemini 3 models (Pro, Flash, Pro Image) with 1M token context; legacy Gemini 2.x and 1.5 models are deprecated Supports text generation, image/audio/video understanding, function calling, structured JSON output, code execution, context caching, and embeddings Official SDKs available: google-genai (Python),...
gemini-interactions-api
google-gemini
Unified interface for Gemini models and agents with server-side state, streaming, and tool orchestration. Supports multiple current models (gemini-3-flash-preview, gemini-3-pro-preview, gemini-2.5-flash/pro) and the Deep Research agent; automatically substitute deprecated model IDs with current alternatives Offload conversation history to the server via previous_interaction_id for stateful multi-turn interactions without manual history management Built-in tool orchestration including...
deliver
google-gemini
Posts a condensed version of the briefing to a Google Chat or Slack incoming webhook, so the daily run delivers itself — skips silently when no webhook is…