AGENT SKILL

content-strategy

Plan a topic map. OmniSkill Social crew — defines audience and goal, builds topic clusters from real demand, and turns them into a sustainable calendar.

License 许可
MIT
所属仓库
IPedrax/OmniSkill
最近更新
2026-10-14
Stars
★ 9
Forks
0

01 前言 Front Matter

name
content-strategy
description
Plan a topic map. OmniSkill Social crew — defines audience and goal, builds topic clusters from real demand, and turns them into a sustainable calendar.
disable-model-invocation
true

03 SKILL.md 正文

Content Strategy

Plans your topic map · OmniSkill Social crew

Deciding what to make, before making it. Run this first for anything ongoing — publishing without a map produces volume, not compounding authority.

1. Define the job

Who — one primary audience, specifically. Their role, their problem, where they already look for answers.

What outcome — organic acquisition, sales enablement, category authority, retention, or hiring. These produce genuinely different content, and a plan optimising for all of them optimises for none.

How success is measured — pick the metric now. Qualified signups, assisted pipeline, branded search volume. "Engagement" is not a business outcome.

2. Build topic clusters

The unit of strategy is the cluster, not the post: one pillar covering a broad topic comprehensively, plus cluster pieces covering sub-topics in depth, all interlinked.

This works because it demonstrates real coverage of a subject rather than isolated pages, and because internal links concentrate authority on the pillar.

Choose 3–5 clusters. More than that spreads thin across every one.

For each cluster:

  • The pillar topic and target query
  • 5–15 cluster sub-topics from real search demand and real customer questions
  • How they interlink — clusters point to the pillar, pillar points to clusters

3. Source topics from evidence

In rough order of value:

  1. Customer questions — sales calls, support tickets, onboarding confusion. Highest intent, and nobody else has this data.
  2. Search demand — keyword research for volume and difficulty. Start where difficulty is winnable, not where volume is largest.
  3. Competitor gaps — what ranks that could be done substantially better, and what nobody has covered.
  4. Community discussion — where the audience already argues about the category.
  5. Original data — proprietary numbers, surveys, benchmarks. The most defensible content there is, because it cannot be copied.

Rank by business relevance × winnability. High-volume terms owned by entrenched publishers are not winnable on a new site, and chasing them wastes the whole quarter.

4. Match format to intent

Informational → guide or explainer. Commercial → comparison or alternatives. Transactional → product or landing page. Navigational → docs.

Publishing a product page against an informational query is the most common structural mismatch, and no promotion fixes it. See /omniskill:seo-audit.

5. Build a sustainable calendar

Set cadence at what can be sustained for a year, not what looks impressive for a month. One strong piece weekly beats five weekly for three weeks and then nothing.

Sequence by dependency: pillar first, then its clusters, then the distribution posts. Reserve roughly 20% of capacity for reactive and opportunistic work.

Plan distribution with creation, not after. A piece with no distribution plan is a piece nobody reads. See /omniskill:social.

6. Maintain what exists

Most teams under-invest here. Updating a piece that already ranks usually beats writing a new one — the compounding is in the refresh, not the volume.

Schedule quarterly review: refresh what is decaying, consolidate pieces competing for the same query, prune what has never earned traffic.

7. Measure

Per cluster, not per post — clusters compound and single posts are noisy. Track organic traffic, rankings for target queries, conversions, and assisted pipeline.

Give it two to three months before judging. Content is a slow instrument and killing a cluster at week four is the most common self-inflicted failure.

Output

Audience and goal statement, 3–5 clusters with pillars and sub-topics, format and intent per piece, a calendar with owners, a distribution plan, and the measurement framework.

Checks before delivering

  • One primary audience, named specifically.
  • Success metric is a business metric.
  • Topics sourced from evidence, not brainstormed.
  • Difficulty is winnable at current authority.
  • Cadence is honestly sustainable.
  • Distribution and maintenance are planned, not assumed.

Feeds /omniskill:pillar-content, /omniskill:social, and /omniskill:email-sequences.

04 同仓库更多 Skills