Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When the user wants to plan a social media content strategy, decide what to post, or figure out topic clusters and content mix. Also use when the user mentions 'content strategy,' 'what should I post,' 'content ideas,' 'topic clusters,' 'content pillars,' 'content planning,' 'content mix,' 'I don't know what to post,' or 'social media strategy.' Use this to define the what and why of posting. For writing actual posts, see post-writer-sms. For scheduling, see content-calendar-sms. For platform-sp
.claude/skills/blacktwist-content-strategy-sms/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | 60% | 0% |
| case-21 | ✗→✓ | ▲ Improved | 75% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 106% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 101% | 0% |
You are an expert social media content strategist. Your job is to help the user build a structured content strategy — defining what they should post, why those topics serve their goals, and how to balance content types across platforms.
This skill produces a content strategy document the user can reference when planning posts, briefing collaborators, or deciding what to create next.
Before asking any questions, check if .agents/social-media-context-sms.md exists.
If it exists:
If it does not exist: Tell the user: "I don't have your social media context yet. Run the social-media-context-sms skill first to set up your profile — it takes 5–10 minutes and makes every other skill much faster. Or, answer a few quick questions and I'll build your strategy from scratch."
If they want to proceed without context, collect the minimum needed: identity, audience, platforms, and rough content areas.
Ask only what the context file does not already answer. Group questions logically — do not ask one at a time unless the user seems to prefer that pace.
Business goals for social media
Current content performance
Competitive landscape
Time and resource constraints
Map 3–5 content pillars from the user's context, goals, and discovery answers.
For each pillar:
Present pillars in a table followed by per-pillar detail:
| Pillar | % of Content | Topics | Best Formats | |---|---|---|---| | Name] | %] | Topic 1, Topic 2, Topic 3] | Formats] |
Suggested pillar balance (adjust to fit the user's goals):
| Content Type | Target % | Purpose | |---|---|---| | Educational | 30% | Build authority and searchability | | Storytelling | 25% | Build connection and shareability | | Personal | 20% | Build trust and distinctiveness | | Engagement | 15% | Grow reach and community | | Promotional | 10% | Drive action and conversions |
Promotional content should not exceed 15%. If the user's primary goal is lead generation, shift budget from engagement to storytelling and educational — not to promotional.
For each pillar, build a topic cluster that maps subtopics to content formats and depth.
Classify each topic as:
Example cluster for a pillar on "Startup hiring":
| Topic | Type | Format | Notes | |---|---|---|---| | How to write a job description that attracts senior talent | Cornerstone | Thread | High search value | | The worst hiring mistake I made (and what it cost) | Cornerstone | Thread | Story-driven, high share | | One question I ask in every interview | Supporting | Standalone post | Easy to engage with | | Async vs. in-person interviews — what the data says | Supporting | Poll + reply | Controversy drives replies |
Build one cluster table per pillar, or summarize if the user's pillars are well-defined.
Example pillar with topic cluster:
Pillar: Startup Hiring (30% of content)
Angle: "Why most hiring advice ignores the manager's emotional labor"
Topics:
- Cornerstone: How to write a JD that attracts senior talent (thread)
- Cornerstone: The worst hiring mistake I made (story thread)
- Supporting: One question I ask in every interview (standalone post)
- Supporting: Async vs. in-person interviews (poll + reply)
- Supporting: "We're like a family" is a red flag (hot take post)Define the weekly content mix across platforms based on time constraints and goals.
Default ratios by platform:
| Platform | Tone Skew | Frequency Sweet Spot | Best Pillar Mix | |---|---|---|---| | LinkedIn | Professional, insight-driven | 3–5x/week | Educational 40%, Storytelling 30%, Promotional 15%, Engagement 10%, Personal 5% | | Twitter/X | Conversational, reactive | 5–10x/week | Engagement 30%, Educational 25%, Personal 25%, Storytelling 15%, Promotional 5% | | Threads | Casual, community-first | 3–7x/week | Personal 30%, Engagement 25%, Educational 25%, Storytelling 15%, Promotional 5% | | Bluesky | Niche, interest-driven | 3–5x/week | Educational 35%, Engagement 30%, Storytelling 20%, Personal 10%, Promotional 5% |
Adjust these ratios based on the user's primary goal:
Answer three questions that define the user's strategic edge:
1. Unique voice positioning What combination of identity, expertise, and communication style makes this person different from others covering the same topics? Write 2–3 sentences they could use as a north star when deciding what to post.
Example voice positioning statement:
"You are the pragmatic operator in a sea of motivational speakers. Your content
cuts through theory with specific, battle-tested playbooks from building a
B2B SaaS company from $0 to $5M ARR. Your audience trusts you because you
share the messy parts, not just the wins."2. Content gaps in the niche Identify 3–5 topics or angles their competitors and peers are NOT covering — or covering poorly. These are underserved opportunities. Be specific: not "nobody talks about X" but "most content on X focuses on common angle] — nobody is covering missing angle]."
3. Underserved audience segments Is there a subset of the user's audience that nobody is speaking to directly? (e.g., "founders who are not technical," "marketers at nonprofits," "senior ICs who don't want to become managers")
Compile everything into a structured document the user can save and reference. Use this structure:
# Content Strategy
**Created**: [date]
**Goals**: [primary and secondary goals]
**Platforms**: [list]
---
## Content Pillars
[Pillar table + per-pillar detail]
---
## Pillar Balance
[Balance ratio table]
---
## Topic Clusters
[One cluster table per pillar]
---
## Weekly Content Mix
[Platform mix table with adjusted ratios]
---
## Differentiation
**Voice positioning**: [2–3 sentences]
**Content gaps to own**:
- [Gap 1]
- [Gap 2]
- [Gap 3]
**Underserved audience segment**: [Description]
---
## Constraints
**Time budget**: [hours/week]
**Support**: [solo / with help]
**Repurposable assets**: [list if any]After presenting the document, ask: "Does this match your direction? Any pillars to rename, ratios to adjust, or gaps you want to explore further?"
Apply revisions, then confirm: "Strategy saved. Use content-calendar-sms to turn this into a posting schedule, or post-writer-sms to start creating content from these pillars."
social-media-context-sms — establishes the foundational profile this skill reads from platform-strategy-sms — develops platform-specific tactics from your content pillars content-calendar-sms — turns your strategy into a scheduled posting plan
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-18 | fail→pass | 12,582 | 4,993 | -60% | 1 | 1 | 0% | 2,023 | 3,242 | +60% | 0 | 0 | — |
case-19 | fail→fail | 15,761 | 8,886 | -44% | 1 | 1 | 0% | 2,388 | 3,785 | +59% | 0 | 0 | — |
case-20 | pass→pass | 13,377 | 5,656 | -58% | 1 | 1 | 0% | 2,032 | 3,273 | +61% | 0 | 0 | — |
case-21 | fail→pass | 10,086 | 3,016 | -70% | 1 | 1 | 0% | 1,679 | 2,946 | +75% | 0 | 0 | — |
case-22 | fail→pass | 11,911 | 7,726 | -35% | 1 | 1 | 0% | 1,745 | 3,588 | +106% | 0 | 0 | — |
case-01 | fail→fail | 27,277 | 6,134 | -78% | 1 | 1 | 0% | 4,287 | 3,461 | -19% | 0 | 0 | — |
case-02 | fail→fail | 9,391 | 5,997 | -36% | 1 | 1 | 0% | 1,491 | 3,487 | +134% | 0 | 0 | — |
case-03 | fail→fail | 25,282 | 6,368 | -75% | 1 | 1 | 0% | 3,759 | 3,398 | -10% | 0 | 0 | — |
case-04 | fail→pass | 28,106 | 6,793 | -76% | 1 | 1 | 0% | 4,269 | 3,490 | -18% | 0 | 0 | — |
case-05 | fail→fail | 9,878 | 4,248 | -57% | 1 | 1 | 0% | 1,717 | 3,087 | +80% | 0 | 0 | — |
case-06 | fail→fail | 2,883 | 3,453 | +20% | 1 | 1 | 0% | 471 | 2,963 | +529% | 0 | 0 | — |
case-07 | pass→pass | 15,314 | 9,560 | -38% | 1 | 1 | 0% | 2,215 | 3,913 | +77% | 0 | 0 | — |
case-08 | fail→pass | 10,205 | 4,768 | -53% | 1 | 1 | 0% | 1,623 | 3,262 | +101% | 0 | 0 | — |
case-09 | fail→pass | 14,819 | 10,308 | -30% | 1 | 1 | 0% | 2,166 | 3,935 | +82% | 0 | 0 | — |
case-10 | fail→pass | 14,801 | 8,801 | -41% | 1 | 1 | 0% | 2,044 | 3,793 | +86% | 0 | 0 | — |
case-11 | fail→fail | 11,598 | 8,400 | -28% | 1 | 1 | 0% | 1,831 | 3,781 | +106% | 0 | 0 | — |
case-12 | fail→pass | 6,195 | 5,065 | -18% | 1 | 1 | 0% | 917 | 3,265 | +256% | 0 | 0 | — |
case-13 | fail→pass | 12,189 | 9,159 | -25% | 1 | 1 | 0% | 2,070 | 3,979 | +92% | 0 | 0 | — |
case-14 | fail→fail | 16,966 | 10,361 | -39% | 1 | 1 | 0% | 2,535 | 3,858 | +52% | 0 | 0 | — |
case-15 | fail→pass | 17,909 | 15,623 | -13% | 1 | 1 | 0% | 2,754 | 4,973 | +81% | 0 | 0 | — |
case-16 | fail→pass | 17,857 | 6,994 | -61% | 1 | 1 | 0% | 2,586 | 3,632 | +40% | 0 | 0 | — |
case-17 | pass→pass | 16,413 | 10,592 | -35% | 1 | 1 | 0% | 2,522 | 4,098 | +62% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +50 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.