OpenSkillIndex← back to search
claude-codeopenclawautomation

Skills news-collect

Run a repo-local scan through `skills/news-aggregator-skill`, then normalize the result into a local report contract for the collect layer. Use when Codex needs a broad overseas or domestic media scan, wants a repeatable `news-report.md`, or needs raw JSON archived under `content-production/inbox/raw/news/`.

install
source · Clone the upstream repo
git clone https://github.com/openclaw/skills
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/openclaw/skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/abigale-cyber/news-collect" ~/.claude/skills/openclaw-skills-news-collect && rm -rf "$T"
OpenClaw · Install into ~/.openclaw/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/openclaw/skills "$T" && mkdir -p ~/.openclaw/skills && cp -r "$T/skills/abigale-cyber/news-collect" ~/.openclaw/skills/openclaw-skills-news-collect && rm -rf "$T"
manifest: skills/abigale-cyber/news-collect/SKILL.md
tags
#news-aggregation#data-processing#report-generation#markdown#json#automation
source content

News Collect

Use the repo-local 

news-aggregator-skill
as a dependency, but always write outputs using this project's collect-layer contract. The report now includes lightweight写作决策字段 for recommended topics.

Quick Start

Run the default command:

.venv/bin/python -m skill_runtime.cli run-skill news-collect --input content-production/inbox/20260405-ai-news-request.md

Prepare Input

Pass a markdown request file with YAML frontmatter.

Supported fields:

  • profile
    : 
    mixed_daily
    / 
    global_latest
    / 
    global_ai
    / 
    cn_media
    / 
    custom
  • sources
    : only used when 
    profile=custom
  • keyword
  • limit
  • deep
  • title

Example:

---
profile: mixed_daily
keyword: AI,Agent
limit: 8
deep: true
title: AI 每日资讯扫描
---

补充说明：更关注能延展成公众号观点文的题目。

Follow Collection Workflow

  1. Read the request frontmatter and resolve the source profile.
  2. Call 
    skills/news-aggregator-skill/scripts/fetch_news.py
    with 
    --no-save
    .
  3. Save the raw JSON to the local inbox raw directory.
  4. Rewrite the result into a stable markdown report for人工选题与后续深研。
  5. For recommended items, also emit写作价值判断、推荐切口、推荐框架与标题方向。

Write Output

Write the report to:

content-production/inbox/YYYYMMDD-<slug>-news-report.md

Write the raw JSON to:

content-production/inbox/raw/news/YYYY-MM-DD/<slug>.json

Respect Constraints

  • Only read the repo-local dependency under 
    skills/news-aggregator-skill/
  • Do not save outputs back into the vendor skill directory
  • Keep the wrapper contract stable even if the vendor skill adds new sources later

Read Related Files

  • Runtime entry: 
    skill_runtime/engine.py
  • Wrapper runtime: 
    skills/news-collect/runtime.py
  • Vendor dependency: 
    skills/news-aggregator-skill/
  • Data contract: 
    docs/data-contracts.md