Tracks RSS news hotspots and generates verification packs and editorial plans.
npx skills add https://github.com/openakita/openakita --skill media-strategy
Please help me install this Agent Skill. Skill: media-strategy Source: https://github.com/openakita/openakita/tree/main/plugins/media-strategy Command: npx skills add https://github.com/openakita/openakita --skill media-strategy
Editors and content teams waste hours manually scanning dozens of news sites and RSS feeds to find worthwhile topics, verify sources, and draft coverage plans. This Skill aggregates multi-source news signals, ranks hotspots by cross-source coverage and source authority, and produces verification checklists and editorial plans through the Media Strategy plugin.
Ask the assistant to recommend today's top 5 media topics from the taiwan package using media_strategy_top_topics, then pick an article_id to generate a verification pack or editorial plan.
High-intent search queries and answers about installing and using this skill.
Call media_strategy_top_topics with a package_id, time window, and limit. It clusters articles across sources and ranks them by multi-source coverage plus authority weighting, returning compact title, link, source count, and risk level per topic.
Use media_strategy_verify_pack with the article IDs and a topic. It generates a verification checklist listing existing sources, missing evidence, and next steps, and flags high-risk items as leads requiring manual review rather than confirmed facts.
Yes, use media_strategy_add_feed with the feed URL and target package. The tool enforces URL safety validation, rejecting localhost, private IPs, and non-http/https schemes, so feeds must be publicly reachable.
The plugin includes built-in HTML listing parsers for sites like taiwan.cn and fjsen.com, with heuristic anchor fallback when selectors fail. For React-rendered sites like ifeng, you can self-host RSSHub and add its route as a custom feed.
Empty results usually mean sources have not been ingested yet or the time window is too narrow. Run media_strategy_ingest first, then widen since_hours to 72 or 168, or lower min_coverage from 2 to 1.
No. It first applies rule-based clustering to select the top N topic clusters, then sends only 3-5 evidence items per cluster to the host LLM in one batch, keeping token cost controlled.
BruceTyndallValidate deduplication guards in rss-to-signals and feed-orchestrator ingestion paths.
SpaceZephyrConvert commercial briefs into structured content plans and sponsored articles.
TereseSongAutomate WeChat Official Account article creation from research to publication.
orthogonal-shAggregate real-time social mentions of brands and topics across platforms.
