briefing

작성자: google-gemini

Formats the standalone daily briefing — curated picks with why-it-matters lines, community buzz, the trend read across runs, and the taste note that keeps the…

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

Briefing Skill

Use this skill to end every sweep. The briefing is the run's product: it must be fully understandable by someone who reads nothing else — no "as I mentioned", no open questions, no references to the conversation.

Instructions for the Agent

Write .agents/workspace/briefings/<YYYY-MM-DD-HHMM>.md (UTC, from date -u). Create the briefings/ directory if needed.

[!IMPORTANT] Exact structure required. The delivery step parses this file to build the chat message, so follow the heading structure precisely:

  • the title is a single # line,
  • each section is a ## line,
  • every item — in every section — is a ### [headline] — [source] line, followed by its why-it-matters sentence, followed by a link line [Source Link](url).

Do NOT use tables, bullet lists for items, or any other layout for the curated items. The ###-heading-plus-link shape is what makes each headline render as a clean clickable link. (Trends, Taste Note, and Data Health are prose/bullets, not items — they have no ### entries.)

Use this exact layout:

# Daily Briefing — [date] — [reader's topics, comma-separated]

## Top Stories
### [headline] — [outlet]
[1-2 sentences: why this matters to THIS reader, tied to their topics or
stated preferences. Mention breadth if multiple outlets covered it.]
[Source Link](url)

### [headline] — [outlet]
[why-it-matters sentence]
[Source Link](url)

## Discoveries
### [headline] — [outlet]
[one line on why this genre-feed item outside the explicit topics earned a slot]
[Source Link](url)

## Community Buzz
### [paraphrased comment/story hook] — Hacker News, [n] points
[One line on what the discussion adds beyond the news coverage.]
[Source Link](url)

## Worth a Longer Read
### [title] — Medium (by [author])
[one line on the depth it offers; member-only stories may be paywalled at the link]
[Source Link](url)

## Trends
[2-4 sentences computed from curation_history.csv: which topics are heating
up or cooling across sweeps, with the numbers cited
("Gemini API: 4 → 9 → 15 items over the last three sweeps").
First run: "First sweep — trend data starts today."]

## Taste Note
[1-3 sentences: which stored preferences shaped today's selection, any
profile change made since the last run, and — at most once per week — one
question-free observation the reader could correct
(e.g. "I've been down-ranking funding news per your 2026-07-15 note.").]

## Data Health
- Fetched: [new_count] new items ([x] news, [y] community, [z] longform) · History rows: [line count of curation_history.csv]
- [Source errors from new_items.json verbatim, or "All sources healthy."]

Rules:

  1. Omit an entire section (its ## header included) when nothing qualified — never emit an empty section or a placeholder item.
  2. Every item links to its source. Headlines may be lightly trimmed but never rewritten into something the outlet did not say.
  3. Community items are paraphrased (≤ 25 words) — never reproduce full comment text.
  4. Every number is derived, never guessed: Fetched counts come from new_items.json (new_count and the per-item source field), and History rows from the line count of curation_history.csv. If you cannot derive a number, omit it rather than estimate.
  5. The Taste Note never asks a question — scheduled runs have no one to answer. It states, in declarative form, what the reader could correct in chat later.
  6. Keep it plain and factual — no meta-commentary about cognitive load, utility, or attention; just the news and why it matters.

google-gemini의 다른 스킬

greeter
google-gemini
친절한 인사 스킬
official
code-reviewer
google-gemini
로컬 변경 사항과 원격 풀 리퀘스트에 대한 자동화된 코드 리뷰로, 정확성, 유지보수성, 보안 측면에서 구조화된 분석을 제공합니다. 로컬 파일 시스템 변경 사항(스테이징 및 언스테이징)과 원격 PR(번호 또는 URL 기준)을 모두 지원하며, 자동 GitHub CLI 체크아웃을 수행합니다. 정확성, 유지보수성, 가독성, 효율성, 보안, 엣지 케이스 처리, 테스트 커버리지의 일곱 가지 차원에서 코드를 분석합니다. 선택적으로 사전 검증 제품군(예: npm run preflight)을 실행하여 문제를 사전에 파악합니다.
official
review-duplication
google-gemini
코드 리뷰 중에 이 스킬을 사용하여 코드베이스에서 중복된 기능, 재발명된 바퀴, 또는 기존 코드를 재사용하지 않은 부분을 사전에 조사하세요.
official
reconciliation
google-gemini
로드된 비용을 사전 파싱된 인보이스 데이터베이스와 대조하여 금액 불일치, 누락된 인보이스, 판매자 불일치 등의 차이점을 플래그 지정합니다…
official
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…
official
async-pr-review
google-gemini
사용자가 비동기 PR 리뷰를 시작하거나, PR에 대한 백그라운드 검사를 실행하거나, 이전에 시작한 비동기 PR의 상태를 확인하려 할 때 이 스킬을 트리거하세요.
official
ci
google-gemini
Gemini CLI를 위한 고성능, 빠른 실패(fail-fast)를 제공하는 특화된 스킬
official
critique
google-gemini
저장소 스크립트와 GitHub Actions 워크플로우를 감사하고 수정하여 기술적 견고성과 보안을 보장하는 전문성.
official