一键导入
synthesize-week
Merge the weekly academia, industry, and community blocks into one structured NEWSLY digest. Use this skill after source gathering is complete.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Merge the weekly academia, industry, and community blocks into one structured NEWSLY digest. Use this skill after source gathering is complete.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Quick-start alias for the NEWSLY pipeline. Trigger this skill when the user says "get news", "get ai news", "run the news pipeline", or wants the simplest possible entrypoint. This skill immediately hands off to `ai_news`.
Master orchestrator for the NEWSLY AI weekly and monthly digest pipeline. Trigger this skill when the user asks for "ai news", "run ai news", "weekly digest", "monthly digest", or any request for a structured AI news roundup. Use this skill as the default entrypoint unless the user explicitly asks to run one sub-skill in isolation.
Gather recent academic AI and ML developments from primary research sources. Use this skill for the research layer of the NEWSLY weekly pipeline. It returns a single structured markdown block covering the most relevant papers, recurring methodological themes, and one under-discussed paper worth closer attention.
Gather recent AI industry developments from model labs, product launches, official API updates, and important open-source tooling releases. Use this skill for the industry layer of the NEWSLY weekly pipeline.
Gather the community pulse around AI from forums, social discussion, and technical content creators. Use this skill for the community layer of the NEWSLY weekly pipeline.
Merge four weekly NEWSLY digests into one higher-order monthly report. Use this skill when all weekly files for a month exist or when the user explicitly forces a monthly synthesis.
| name | synthesize_week |
| description | Merge the weekly academia, industry, and community blocks into one structured NEWSLY digest. Use this skill after source gathering is complete. |
You receive three pre-fetched blocks:
ACADEMIA_BLOCKINDUSTRY_BLOCKVIRAL_BLOCKYour job is to turn them into one coherent weekly digest for an applied AI reader who cares about research rigor and practical relevance.
Before writing, identify overlaps across the three blocks.
Rules:
also discussed by the community when usefulIdentify 2 or 3 themes that span multiple blocks.
Good themes connect at least two of:
Return the full markdown file in exactly this structure:
---
week: [N]
month: [MONTH]
year: [YEAR]
date_range: "[MONTH] [start_day]-[end_day], [YEAR]"
generated_at: "[ISO timestamp]"
sources_count:
academia: [N papers]
industry: [N items]
viral: [N items]
---
# AI Digest - [MONTH] [YEAR], Week [N]
---
## Academia & Research
[Insert cleaned ACADEMIA_BLOCK content]
---
## Industry & Big Tech
[Insert cleaned INDUSTRY_BLOCK content]
---
## Viral & Community Pulse
[Insert cleaned VIRAL_BLOCK content]
---
## Cross-Cutting Themes This Week
[Theme 1 - 2-3 sentences]
[Theme 2 - 2-3 sentences]
[Theme 3 - optional]
---
## Research Lens
*What this week means for academic work specifically.*
[3-5 sentences.]
**One actionable call:** [1 specific sentence.]
---
## Lab Lens
*What this week means for practical lab work.*
[3-5 sentences.]
**One actionable call:** [1 specific sentence.]
---
## Analyst's Take
*Opinionated synthesis after the neutral sections.*
### What's Actually Interesting
[2-3 sentences distinguishing real advances from loud but thin coverage.]
### What Everyone Missed
[2-3 sentences on an underweighted development.]
### The Slow Burn to Watch
[1-2 sentences on a trend building over multiple weeks.]
Do not split the final writing across parallel workers. Upstream source gathering may be parallelized, but this final synthesis should be produced by one agent so the digest has one coherent editorial voice.