Plan First, Write Faster: The writing-article-plan Skill

Where this fits

You’ve just finished a brainstorming session with your AI agent. The topic brief is approved: a 1,200-word piece on "automating customer support with LLMs." Now what? If you’re like most developers, you prompt the agent to "write the article." It produces a wall of text that drifts off-topic, repeats points, and needs heavy editing. The writing-article-plan skill solves this by forcing a structured outline before any drafting. It fits right after topic approval and before the actual writing phase—turning a vague instruction into a precise, section-by-section execution plan.

Why agents benefit

• Breaks down large tasks: Each section is 150–400 words, so the agent can write one focused piece at a time without losing coherence.
• Reduces context switching: The plan captures all key points, transitions, and word counts per section, so the agent doesn't need to re-read the brief or re-plan mid-writing.
• Enforces consistency: The header template and section granularity ensure every article has the same structure—introduction, three body arguments, conclusion—making output predictable.
• Explicit research needs: The plan lists what evidence or sources are required before writing, so the agent can gather them in advance instead of pausing mid-draft.
• Saves memory and tokens: By writing to a plan file, the agent can reference it during drafting without holding the entire outline in context.

Practical scenarios

Scenario 1: Solo developer juggling multiple content pieces You manage a technical blog and a newsletter. With writing-article-plan, you feed each approved brief into the skill, get a structured outline, and then delegate each section to your agent as a separate task. You can review the plan before any words are written, catching structural issues early.

Scenario 2: Team collaboration with version control The plan is saved to docs/writing/YYYY-MM-DD-<topic>-plan.md. Your team can review the outline, comment on specific sections, and approve changes before the agent proceeds. This turns AI writing into a collaborative, reviewable process.

Scenario 3: Long-form content that needs research For a 3,000-word in-depth guide, the skill forces you to specify research needs per section. You can instruct the agent to gather citations, stats, or examples for each body section before writing, ensuring the final article is well-supported.

Add it to your agent workflow

Install the skill with:

npx skills add binbinao/document-superpowers --skill writing-article-plan

Then use it in your agent prompt like this:

I have an approved topic brief: "How to Build a RAG Pipeline in 2025".
Use the writing-article-plan skill to create a detailed outline.
Save the plan to docs/writing/2025-03-15-rag-pipeline-plan.md.

The agent will announce it's using the skill, generate the plan, and save it for your review.

Compare before adopting

Before committing, compare this skill with:

• Quality signals: Check the GitHub repo (44 stars) for recent commits, issues, and whether the skill is actively maintained.
• Alternatives: Other outline-generation skills exist, but this one is tightly scoped to article planning and integrates with the document-superpowers collection.
• Workflow fit: Ensure your agent can handle the file-saving step and that you have a process for reviewing plans. If you usually write in a single prompt, this adds an extra step—but it often saves time in editing.

Why it is worth tracking

This skill addresses a common pain point: AI-generated articles that lack structure. Its community momentum is modest but steady, and the author has built a suite of document-focused skills. Track it if you're building content workflows where quality and consistency matter. Evaluate it after a few uses to see if the planning overhead pays off in reduced revision cycles. For a curated list of similar skills, check out OpenAgentSkill for updates and community feedback.