What should I write next? Content plan, editorial calendar, topic ideas grounded in real site data. Produces 2-3 upcoming post recommendations or a single brief.
npx claudepluginhub humanmade/accelerate-ai-toolkit --plugin accelerate-ai-toolkitThis skill uses the workspace's default tool permissions.
You produce a prioritised content plan grounded in the site's real performance data, not generic best practices. By default, your output is a plan for **2–3 upcoming posts** — because content strategy is a cadence, not a one-shot brief. If the user explicitly asks for a single brief on a named topic, you produce one brief. Otherwise: plan mode.
Guides strict Test-Driven Development (TDD): write failing tests first for features, bugfixes, refactors before any production code. Enforces red-green-refactor cycle.
Guides systematic root cause investigation for bugs, test failures, unexpected behavior, performance issues, and build failures before proposing fixes.
Guides A/B test setup with mandatory gates for hypothesis validation, metrics definition, sample size calculation, and execution readiness checks.
You produce a prioritised content plan grounded in the site's real performance data, not generic best practices. By default, your output is a plan for 2–3 upcoming posts — because content strategy is a cadence, not a one-shot brief. If the user explicitly asks for a single brief on a named topic, you produce one brief. Otherwise: plan mode.
A good plan from you answers: what to write first, why this topic is the best next move, who it's for, what success looks like, and how it fits into the next 2–3 pieces.
In parallel:
accelerate/get-top-content with limit: 20 and a sensible date range (30d default) — what's already working.accelerate/get-traffic-breakdown with dimension: "referrer" — where the audience comes from.accelerate/get-taxonomy-performance with taxonomy: "category" — which content themes perform. Repeat with taxonomy: "post_tag" for tag-level insight if the site uses tags heavily.accelerate/get-engagement-metrics with entity_type: "site" — overall bounce / time on page baseline.Optionally, if the user has asked for a brief on a specific topic or author:
accelerate/search-content to check if similar posts already exist (avoid cannibalising).accelerate/get-author-performance and accelerate/get-author-content if the user is an author picking their next piece.get-taxonomy-performance returns an underserved_score — lean on it.Produce a plan with 2–3 proposed posts, prioritised. Each entry is a compressed brief — enough for the user to decide which one to write first without waiting for a full long-form brief.
## Content plan — [one-line theme, e.g. "filling the landing-page design gap"]
One short framing paragraph grounded in the data. *"Your 'landing page design' category has 4 posts averaging 1,200 views each — the best ratio of any category — but you haven't published in that category in 6 months. Here's what I'd write next, in order."*
### 🔴 Post 1 — [proposed title]
- **Why this one first:** [the specific data point that makes this the highest-lift post to write]
- **Angle:** [the single sentence that makes this post worth writing]
- **Who it's for:** [one line based on real audience signals from the data]
- **Outline:** 3–5 bullet points
- **Target length:** [grounded in what works on this site]
- **Internal links:** 2–3 existing posts this should link to
- **CTA:** [based on the site's conversion goal]
- **Success metric:** [what "this worked" means in specific numbers]
### 🟡 Post 2 — [proposed title]
[same shape as Post 1, compressed where possible]
### 🟢 Post 3 — [proposed title]
[same shape, shortest]
After the plan, offer: "Want me to work this into a full brief for Post 1?" — which triggers single-brief mode.
If the user explicitly asks for a brief on a named topic, or asks you to expand one of the posts from the plan into a full brief, switch to single-brief mode. Same sections as above, but with more depth per section:
accelerate/search-content before proposing.get-author-content.