content-refresher
by aaron-he-zhucontent-refresher is an SEO workflow skill that helps diagnose content decay, rule out false positives, and build structured refresh plans for aging articles with dropping rankings or traffic.
This skill scores 84/100, meaning it is a solid directory listing candidate: agents get strong trigger cues, a substantial refresh workflow, and concrete reference material that should reduce guesswork versus a generic prompt. Directory users can reasonably understand what it does and whether it fits SEO/content-refresh work, though they should expect mostly document-based guidance rather than executable tooling.
- Highly triggerable: frontmatter includes many explicit multilingual triggers tied to outdated content, ranking drops, and traffic decline.
- Operationally substantial: SKILL.md is long and structured, with multiple workflow and constraint signals plus code-fenced templates and repo references.
- Good install-decision support: reference files cover decay detection, worked examples, and refresh templates, helping users judge fit before installing.
- No install command or bundled scripts, so execution depends on the agent following documentation rather than running a packaged workflow.
- SEO-tool integration is optional and external, which may leave some users guessing about how much data access is needed for best results.
Overview of content-refresher skill
What content-refresher does
content-refresher is an SEO workflow skill for updating aging articles that have lost freshness, rankings, or traffic. It is built for a specific job: diagnose content decay, decide whether a page is worth refreshing, and produce a structured refresh plan instead of vague “make this better” edits.
Who should use this skill
Best fit:
- SEO writers and content strategists managing older blog libraries
- agencies handling ranking recovery for existing pages
- site owners who see “traffic is dropping” and need a repeatable refresh process
- teams that want a clearer refresh brief before rewriting content
It is less useful if you are creating a net-new article from scratch or debugging technical SEO issues first.
The real job-to-be-done
Most users do not just want rewritten copy. They want to answer:
- is this page actually decaying, or is something else wrong?
- what should be updated first?
- what can stay untouched?
- what new sections, stats, and freshness signals will most likely restore performance?
The content-refresher skill is strongest when you need that decision layer, not just text generation.
What makes content-refresher different
The main differentiator is structure. The skill includes:
- decay detection guidance with severity thresholds
- false-positive checks so you do not blame content for seasonality or technical problems
- refresh planning templates
- a worked example showing the expected output shape
That makes it more useful than a generic prompt for “update this article for SEO,” especially when multiple stakeholders need a defensible plan.
What to inspect before installing
Read these first:
optimize/content-refresher/SKILL.mdoptimize/content-refresher/references/content-decay-signals.mdoptimize/content-refresher/references/refresh-example.mdoptimize/content-refresher/references/refresh-templates.md
If those files match how your team already evaluates declining content, the content-refresher skill is likely a good fit.
How to Use content-refresher skill
Install context for content-refresher
The repository indicates compatibility with Claude Code, skills.sh marketplace, ClawHub, and the Vercel Labs skills ecosystem. No system packages are required. Optional network or MCP access only matters if you want live SEO-tool data during analysis.
If your environment supports GitHub skills directly, install from the repository path that contains optimize/content-refresher.
Start with the right repository files
Do not begin with the full skill body alone. Use this order:
SKILL.mdfor trigger conditions and workflowreferences/content-decay-signals.mdfor prioritization logicreferences/refresh-example.mdto see the expected analysis formatreferences/refresh-templates.mdto turn findings into usable deliverables
This reading path reduces guesswork and shows whether the skill outputs what your team actually needs.
What input content-refresher needs
The content-refresher skill works best when you provide more than a URL and “please update this.” Strong inputs include:
- page URL
- current title and publish date
- last updated date
- target keyword or topic cluster
- current rankings or traffic trend
- known competitors now outranking the page
- article text or section outline
- outdated stats, products, screenshots, or links you already suspect
- business goal: recover rankings, improve conversions, win snippets, or support GEO visibility
Without this context, the skill can still draft a plan, but the recommendations will be more generic.
Turn a rough goal into a usable prompt
Weak prompt:
- “Refresh this article for SEO.”
Stronger prompt:
- “Use the content-refresher skill on this post:
[URL]. Traffic is down 32% YoY and rankings dropped from positions 4-5 to 10-12 for[keyword]. Published in 2023, never updated. I need a refresh plan that identifies what to keep, what to update, what new sections to add, stale statistics to replace, and whether the title should change.”
That prompt gives the skill a trigger, severity, business objective, and output format.
Best prompt pattern for content-refresher usage
A practical structure is:
- state the decay signal
- provide the page context
- define the goal
- request a staged output
Example:
- “Apply content-refresher to this blog post. First, check whether the decline looks like true content decay versus seasonality or technical issues. Second, score refresh priority. Third, produce a refresh plan with section-level actions, updated title options, FAQ opportunities, source gaps, and a prioritized rewrite checklist.”
This aligns closely with the skill’s built-in references.
Suggested workflow for real refresh projects
Use the skill in four passes:
- Diagnosis: confirm decay and rule out false positives
- Planning: identify sections to keep, expand, remove, or add
- Rewrite brief: generate concrete instructions for the writer
- Execution review: compare the revised draft against the refresh plan
This workflow is better than asking for a full rewrite immediately because it prevents unnecessary rewrites and keeps useful material intact.
How to use performance signals correctly
One of the strongest parts of content-refresher for SEO Content is the decay signal framework. The reference file uses thresholds for traffic and ranking decline, plus urgency bands like watch, warning, critical, and emergency.
Use those thresholds as decision support, not absolute truth. Before refreshing, check:
- seasonality with year-over-year comparisons
- known algorithm updates
- crawl/indexing issues
- analytics tracking changes
- product or market changes that altered search intent
If you skip this step, you may refresh the wrong page or solve the wrong problem.
What outputs you should expect
A good content-refresher output should include:
- a quick content quality or trust assessment
- evidence that the page is actually stale
- prioritized refresh actions
- title and structure recommendations
- statistics and links to update
- new sections or FAQ blocks to add
- a rewrite scope that separates “keep” from “replace”
If the output jumps straight into rewritten paragraphs without explaining the refresh logic, ask for the plan first.
When to ask for a full rewrite
Ask for the rewrite only after the refresh plan is sound. Best sequence:
- first prompt: diagnosis + refresh plan
- second prompt: section-by-section rewrite using the approved plan
- third prompt: metadata, FAQ, internal links, and on-page freshness updates
This staged content-refresher usage produces cleaner editorial control than one-shot generation.
Practical tips that improve output quality
- Paste the existing headings so the model can mark keep/update/remove decisions.
- Provide traffic or rank deltas, not just “it dropped.”
- Name the SERP competitors that replaced you.
- Tell the model whether the page should stay evergreen or become year-specific.
- Ask for source-replacement placeholders when live web verification is unavailable.
- Request “minimal necessary changes” if you want ranking recovery without full structural churn.
Misfit cases and adoption blockers
Do not use content-refresher as your first move when:
- the page never ranked in the first place
- the main issue is technical SEO
- search intent has fully changed and the page needs replacement, not refresh
- you need a compliance-reviewed publication workflow with citations automatically validated
The skill is a strong planning framework, but it does not remove the need for factual verification and editorial judgment.
content-refresher skill FAQ
Is content-refresher better than a normal SEO prompt?
Usually yes, if your problem is declining performance on an existing page. A normal prompt may rewrite text, but the content-refresher skill gives you a refresh-specific process: detect decay, avoid false positives, prioritize changes, and structure the update.
Is content-refresher beginner-friendly?
Yes, more than many SEO skills, because the references show thresholds, examples, and templates. Beginners still need to know basic page metrics and how to judge whether a source or stat is current.
Can I use content-refresher without SEO tools?
Yes. You can manually provide traffic, ranking, or qualitative decline signals. The result will be better if you include data from Search Console, analytics, or third-party rank tools, but the skill does not require them to function.
Is content-refresher only for blog posts?
No, but blog and evergreen editorial pages are the clearest use case. It can also help with guides, listicles, comparison pages, and informational landing pages. It is less natural for product pages or highly programmatic content.
What makes content-refresher for SEO Content useful?
Its value is not just “update old text.” It helps decide:
- whether the page is worth saving
- which freshness signals matter
- how much of the original structure should survive
- where new sections, FAQs, and updated evidence can lift relevance
That is especially useful for content teams managing many aging URLs.
When should I not use content-refresher?
Skip it when the page has technical indexing problems, cannibalization issues, or a complete intent mismatch with current SERPs. In those cases, fix architecture or page targeting before running a refresh workflow.
Does the skill generate final publish-ready copy?
It can support that, but its strongest value is upstream: analysis, planning, and rewrite guidance. Treat it as a decision-making and briefing tool first.
How to Improve content-refresher skill
Give content-refresher better evidence
The biggest output upgrade comes from better inputs. Include:
- old and current title
- publication history
- rank and traffic change windows
- top queries affected
- competitor pages now winning
- stale claims, screenshots, examples, or pricing data
- internal links pointing to the page
This lets the skill produce an actionable refresh plan instead of broad advice.
Ask for section-level decisions
A common failure mode is getting generic recommendations like “add more depth.” Prevent that by asking:
- “For each current H2, label keep, update, merge, remove, or replace.”
- “List exact new sections to add and why.”
- “Flag outdated year references, stats, products, and broken links.”
This turns content-refresher into an editorial operations tool, not just an idea generator.
Separate diagnosis from rewriting
If you ask for everything at once, the model may over-rewrite before proving the page should be refreshed. Improve results by running content-refresher in stages:
- decay validation
- refresh plan
- rewrite
- QA against the plan
That sequencing usually improves trust and reduces wasted editing.
Push for source-aware updates
Another common failure mode is fake freshness: changing wording without updating evidence. Ask the skill to produce:
- a list of statistics requiring replacement
- citation placeholders for current sources
- claims needing fact checks
- examples that should be swapped for more recent ones
For SEO content, freshness signals only help if the underlying information is actually current.
Constrain the rewrite to your intent
Tell the skill whether you want:
- conservative refresh for ranking recovery
- major expansion to compete with stronger SERPs
- GEO-oriented FAQ additions
- conversion-focused updates while preserving search intent
Without this instruction, the output may drift into a larger rewrite than you need.
Use the provided references as reusable templates
The strongest way to improve content-refresher usage is to reuse its reference files as your working format:
- decay thresholds from
references/content-decay-signals.md - output shape from
references/refresh-example.md - section planning from
references/refresh-templates.md
That makes team reviews faster because everyone works from the same refresh grammar.
Watch for these common failure modes
- treating seasonal drops as content decay
- updating copy but leaving stale numbers or links
- adding sections that dilute the original search intent
- changing titles aggressively without considering URL/topic continuity
- refreshing pages that should actually be consolidated or replaced
Most poor outcomes come from misdiagnosis, not weak writing.
Iterate after the first output
After the first plan, ask follow-up questions like:
- “Which 3 changes are highest ROI if I can only edit for one hour?”
- “What should I update without changing the URL?”
- “Which additions are most likely to improve snippet capture?”
- “What evidence gaps would block publication?”
These follow-ups make the content-refresher skill more practical for real deadlines.
How to judge whether the refresh plan is good
A strong plan is specific enough that a writer can execute it without guessing. It should clearly state:
- why the page declined
- whether refresh is the right fix
- what stays unchanged
- what must be updated now
- how the revised structure will better match current SERPs
If those answers are missing, run another pass before drafting.
