Content Calendar

Build a data-backed content calendar. Scrapes the user's own blog to understand current coverage, profiles competitor topics, pulls keyword gap data from Ahrefs, optionally layers in GSC/GA4 performance signals, then produces a prioritized list of articles to write over the next quarter.

Setup

Recommended:

• Ahrefs MCP — the official remote MCP server at https://api.ahrefs.com/mcp/mcp. Connect via Claude Code's MCP settings (OAuth, paid Ahrefs plan required). Used here for keyword volume, difficulty, and content-gap data via Keywords Explorer and Site Explorer.
• FIRECRAWL_API_KEY — crawl own blog and competitor sitemaps.

Optional:

• GSC_SERVICE_ACCOUNT_JSON + GSC_SITE_URL — declining pages to refresh, striking-distance queries to push.
• GA4_SERVICE_ACCOUNT_JSON + GA4_PROPERTY_ID — top-performing existing content to cluster around and repurpose.
• SE_RANKING_API_KEY — alternative rankings data source.

Inputs

1. Own blog URL — root or blog index.
2. Competitor list — 3–10 domains (can reuse the list from a prior competitor-analysis run if present at competitors/<date>-raw.json).
3. Calendar horizon — one month, quarter, or half year. Default quarter.
4. Cadence — posts per week or month the user plans to publish.
5. Focus areas — optional thematic pillars (e.g. "sales enablement, RevOps, onboarding"). Inherit from editorial-guidelines.md if the user has listed content pillars there.
6. Include refreshes? — default Y. Reuse GSC declining-pages data if configured.

Use AskUserQuestion for horizon and refreshes Y/N.

Process

1. Inventory the user's own blog

Scrape the user's blog via the Firecrawl → Jina → Playwright chain:

• Pull the sitemap if available; otherwise crawl the blog index.
• For each post, extract: URL, title, H1, publish date, updated date, primary topic/entity cluster, rough word count.

Cluster existing posts into topics using entity overlap. Produce a map of {cluster: [posts]}.

2. Inventory the competitors

Reuse competitors/<latest>-raw.json if present. Otherwise crawl each competitor blog via the same chain and build the same cluster map.

Flag topics where:

• Competitors have ≥3 posts and the user has 0 → white-space topic.
• Competitors have ≥3 posts and the user has 1 → under-covered topic.
• User has ≥3 posts, competitors have 0–1 → user-dominated topic (defend).

3. Keyword gap (Ahrefs MCP)

Call the Site Explorer → content gap tool: own domain vs competitors. Capture keywords all competitors rank for but the user doesn't. Enrich each with volume, KD, and parent topic via Keywords Explorer → overview.

Group keywords into topic clusters. Each cluster becomes a candidate article with:

• Suggested primary keyword (highest-volume with KD in range)
• 5–10 secondary keywords
• Estimated traffic ceiling (sum of volumes × conservative CTR)

4. Performance signals (optional)

GSC:

• Striking distance (pos 8–20): convert each into either a new post or an optimization of an existing one.
• Declining queries: find the landing page, suggest a refresh.
• Impressions but no clicks (high impressions, CTR <1%): meta/title rewrite candidates.

GA4:

• Top 20 pages by engaged sessions in the last 90 days → suggest a cluster expansion around each.
• Pages with high entrance but high exit rate → UX + internal-link fixes.

5. Score and prioritize

For each candidate article, compute a score:

score = (volume × relevance_to_pillars × brand_fit)
      / (kd × effort_estimate)
      × traffic_performance_multiplier

Where:

• volume — Ahrefs monthly search volume
• relevance_to_pillars — 0.5 if off-pillar, 1.0 on-pillar, 1.3 on-pillar + in white-space topic
• brand_fit — 1.0 default, 0.7 if the topic doesn't match the brand's ICP
• kd — Ahrefs difficulty, floored at 5
• effort_estimate — 1 for comparison pages and listicles, 2 for deep how-to, 3 for original-research pieces
• traffic_performance_multiplier — 1.2 if there's GSC striking-distance overlap, 1.0 otherwise

Rank candidates by score. Split into:

• Must-write (top N to fill cadence)
• Nice-to-write (next N)
• Refresh queue (existing posts to update)

6. Sequence the calendar

Assign each must-write entry to a week based on:

• Cadence (e.g. 2 posts/week)
• Topic clustering (don't publish two posts from the same cluster in the same week)
• Seasonality if the user flagged any
• Any known product launches or events the user mentioned

7. Pre-seed briefs

For each must-write entry, create a lightweight seed in briefs/<slug>.stub.md with:

• Primary keyword
• Secondary keywords
• Estimated volume and KD
• Topic cluster
• Suggested title
• One-line angle

The user can later run the content-brief skill against each stub to expand it into a full brief.

8. Write the calendar file

Write to calendar/<horizon>-<YYYY-MM>.md:

# Content Calendar — {horizon}

_Generated: {YYYY-MM-DD}_
_Own blog: {url}_
_Competitors: {n}_
_Cadence: {posts}/week_
_Horizon: {start} → {end}_

## Summary

- **New articles planned:** {n}
- **Refreshes planned:** {n}
- **Estimated traffic ceiling:** {n}/month
- **Top topic clusters to win:** {list}

## Must-write

| Week | Date | Title | Primary kw | Volume | KD | Cluster | Score | Brief stub |
|---|---|---|---|---|---|---|---|---|
| W1 | 2026-04-15 | {title} | {kw} | 2,400 | 32 | {cluster} | 48 | `briefs/{slug}.stub.md` |

## Nice-to-write (backlog)

| Title | Primary kw | Volume | KD | Why it didn't make the cut |
|---|---|---|---|---|

## Refresh queue

| Existing URL | Declining query | Current pos | Potential | Recommended change |
|---|---|---|---|---|

## Cluster coverage map

| Cluster | Existing posts | Planned | Competitor posts | Gap |
|---|---|---|---|---|

## Notes

- White-space topics: {list}
- Threats (user-dominated topics where competitors are gaining): {list}
- Sources used: Ahrefs {Y/N}, Firecrawl {Y/N}, GSC {Y/N}, GA4 {Y/N}

Also save raw scoring data to calendar/<horizon>-<YYYY-MM>.raw.json.

9. Print summary

One block: N articles planned, N refreshes queued, top cluster, path to calendar, path to stubs directory.

Fallbacks

• No Ahrefs MCP: skip keyword scoring; rank candidates by (relevance × competitor_coverage_count). Clearly mark the calendar as "content-only prioritization" in the header.
• No GSC / GA4: skip performance signals; rely entirely on gap analysis. The traffic multiplier becomes 1.0 for everything.
• Competitor blog scrapes time out: work with partial competitor coverage; list which competitors were skipped.
• No editorial guidelines pillars: ask the user directly or derive pillars from the cluster analysis of their existing blog.

Verification

1. Calendar file exists.
2. Must-write table has one row per planned slot in the horizon (no gaps).
3. Every must-write entry has a matching stub file under briefs/.
4. Raw JSON file present.
5. If re-run a month later, diffing the raw files should show what moved from backlog to must-write and which refreshes were completed.