flaqai/Awesome_SEO_Writing_Skill / writer/linkedin-writer
linkedin-writer
VerifiedLinkedIn-native long-form article and newsletter writing workflow for LinkedIn and Google-to-LinkedIn topic discovery, business-depth research, professional thought leadership, evidence-led drafting, final humanization, discussion design, SEO settings, auditing, and publish-ready packaging. Use when creating, outlining, researching, enriching, rewriting, humanizing, auditing, or packaging LinkedIn Articles, LinkedIn newsletters, LinkedIn long-form posts, LinkedIn thought leadership, LinkedIn B2B articles, LinkedIn 长文, LinkedIn 专栏, LinkedIn 话题调研, LinkedIn 商务内容, 去 AI 化编辑, or LinkedIn 发布包.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 23.4 KB |
| references/linkedin-article-template.md | markdown | 15.3 KB |
| references/linkedin-business-depth-and-humanization.md | markdown | 12.1 KB |
| references/linkedin-review-rubric.md | markdown | 12.3 KB |
| references/linkedin-topic-research.md | markdown | 15.0 KB |
| scripts/audit-linkedin-markdown.mjsexecutable | script | 18.5 KB |
references/linkedin-review-rubric.md
View on GitHub# LinkedIn Article Review Rubric
Use this rubric after fact checking and before final packaging. Review the article as a professional reader, not only as a Markdown document.
## Review Output
Save the result as `linkedin-audit.md`:
```markdown
# LinkedIn Article Audit
## Verdict
- Status: PASS / REVISE / BLOCKED
- Target audience:
- Core thesis:
- Strongest element:
- Highest-priority issue:
## Scorecard
| Dimension | Score (1-5) | Evidence | Required change |
|---|---:|---|---|
| Professional relevance | | | |
| Business depth | | | |
| Thesis and originality | | | |
| Evidence and trust | | | |
| Usefulness and application | | | |
| LinkedIn readability | | | |
| Conversation quality | | | |
| Publishing completeness | | | |
## Blocking issues
- ...
## Priority findings
| Priority | Area | Evidence | Impact | Fix |
|---|---|---|---|---|
| High | ... | ... | ... | ... |
## Revision summary
1. ...
## Fact-check summary
- Verified:
- Softened:
- Removed:
- Still limited:
## Humanization integrity check
- Voice target:
- Mechanical patterns revised:
- Generic business abstractions made concrete:
- Template symmetry revised:
- First-person provenance verified:
- Facts and certainty preserved:
- Links, names, disclosure, and topic phrases rechecked:
- AI-detector bypass claimed: no
## Deterministic audit
- Command:
- Result:
- Warnings accepted or fixed:
## Publication status
- Local package only / scheduled / published and verified
```
## Passing Standard
The article passes only when:
- average score is at least 4.0;
- no dimension is below 3;
- all blocking issues are resolved;
- material claims are verified, softened, or removed;
- relevant business decision, stakeholder, economic, implementation, measurement, risk, scenario, and action context has been developed rather than padded;
- final LinkedIn humanization has been completed and integrity-checked;
- the final article and operational publishing pack are separable;
- the deterministic audit has been run and high-impact warnings are resolved or explained.
Revise within the existing evidence up to two rounds. If a central claim remains unsupported, narrow the thesis or remove the claim. Do not give an unsupported draft a passing score because it sounds polished.
## Blocking Issues
Any one of these blocks final delivery as publish-ready:
- invented professional experience, customer story, test, quote, interview, metric, or internal result;
- a materially false or unsupported core claim;
- a headline or opening that promises evidence or outcomes the body does not provide;
- copied phrasing, structure, story, framework, or conclusion from another creator;
- a Google snippet, LinkedIn post, or creator claim presented as verified fact without adequate source verification;
- undisclosed material commercial interest or affiliation;
- harmful professional advice presented without necessary limits;
- wrong audience, product, model, company, policy, or topic;
- broken required link, incorrect attribution, or fabricated citation;
- publishing notes or internal audit text mixed into the public article body;
- a claim that the article was published, indexed, delivered, or performed without verification.
## Score Dimensions
### 1. Professional relevance
**5:** Names a specific role, situation, and decision. The stakes are clear and current evidence supports why the topic matters now.
**4:** Audience and work problem are clear, with minor room for sharper context.
**3:** Generally useful to professionals but too broad by role, industry, seniority, or situation.
**2:** Reads like generic content marketing with weak work relevance.
**1:** No identifiable professional reader or decision.
Check:
- Can the target reader recognize their situation in the first 120 words?
- Does the article solve a work problem rather than merely describe a trend?
- Are adjacent topics relevant to the same reader decision?
- Is “why now” supported rather than asserted?
### 2. Business depth
**5:** Connects the topic to a real decision and develops the relevant stakeholder, economics, implementation, measurement, risk, scenario, and next-action context with evidence and clear ownership.
**4:** Strong business context with one secondary dimension left thin.
**3:** Mentions business implications but stays abstract around owners, costs, metrics, prerequisites, or trade-offs.
**2:** Adds trend facts and management language without explaining how a decision would actually be made or implemented.
**1:** No meaningful business context, or the article hides a material cost, stakeholder, risk, or limitation.
Check:
- Is there a specific decision, trigger, and cost of waiting?
- Are sponsors, users, approvers, blockers, buyers, and post-launch owners visible where relevant?
- Does the article distinguish measurable economics from an ROI assumption?
- Are workflow, integration, adoption, and change-management dependencies concrete?
- Are baseline, outcome metric, guardrail, and review owner named where needed?
- Is the strongest objection treated fairly?
- Is at least one example attributable or explicitly hypothetical?
- Does the reader know which artifact, meeting, pilot, or review should happen next?
### 3. Thesis and originality
**5:** Makes one clear, defensible point that combines evidence and judgment into an original synthesis.
**4:** Thesis is clear and useful, though one section may be familiar or underdeveloped.
**3:** Topic is clear but the article summarizes more than it argues.
**2:** Uses a generic “benefits, challenges, future” structure with little distinction.
**1:** Thesis is missing, contradictory, copied, or replaced by promotion.
Check:
- Can the thesis be stated in one sentence?
- Does each H2 advance it?
- Is the point different from a generic search summary?
- Does the article acknowledge what would weaken or falsify its recommendation?
### 4. Evidence and trust
**5:** Material facts are current, attributable, scoped correctly, and separated from inference. Experience and affiliation boundaries are explicit.
**4:** Claims are reliable with small attribution or recency improvements needed.
**3:** Main facts are sound but evidence is thin or too dependent on secondary sources.
**2:** Multiple vague, stale, or unsupported claims affect the recommendation.
**1:** Core evidence is false, invented, plagiarized, or materially misleading.
Check:
- Are dates, numbers, products, policies, and comparisons verified?
- Does research retain its population, geography, and date?
- Is first-person experience user-provided or otherwise authorized?
- Are vendor claims clearly framed as vendor claims?
- Are Google-discovered LinkedIn results treated as conversation or practitioner evidence rather than automatic factual proof?
- Is disclosure present when recommendation and commercial interest overlap?
### 5. Usefulness and application
**5:** Gives concrete criteria, steps, examples, checks, role implications, and conditions that the reader can apply.
**4:** Provides a strong practical framework with one thin section.
**3:** Contains useful advice but stays abstract in important places.
**2:** Repeats broad recommendations such as “adapt,” “learn,” or “be strategic” without showing how.
**1:** Offers no actionable or decision-useful value.
Check:
- Does every main section add a decision criterion, method, example, or implication?
- Does the reader know what to do differently?
- Are prerequisites, risks, and validation steps included?
- Are recommendations conditional by role or scenario when needed?
### 6. LinkedIn readability
**5:** Strong opening, clear TL;DR when appropriate, descriptive headings, varied short-to-medium paragraphs, useful formatting, and a natural pace.
**4:** Easy to scan with minor density or repetition.
**3:** Readable but one or more sections feel blog-like, dense, overly uniform, or list-heavy.
**2:** Long blocks, repeated heading patterns, mechanical transitions, or excessive promotion create friction.
**1:** Difficult to read or structurally broken.
Check:
- Do the first 120 words establish tension, thesis, and promise?
- Can headings be understood in isolation?
- Are paragraphs usually 1-4 sentences?
- Does bold, list, quote, table, and image use help rather than decorate?
- Does humanization preserve a credible professional voice?
- Did the final pass remove generic abstraction, synthetic LinkedIn performance, and excessive template symmetry without manufacturing personality?
### 7. Conversation quality
**5:** Surfaces a real tension, treats disagreement fairly, and ends with a precise question qualified readers can answer usefully.
**4:** Invites relevant discussion with a minor opportunity to sharpen the question.
**3:** Has a relevant question but limited tension or specificity.
**2:** Uses generic engagement prompts, artificial controversy, or one-sided certainty.
**1:** Uses spammy engagement bait, irrelevant tagging, or no discussion path.
Check:
- Is the strongest counterargument represented accurately?
- Can a reader disagree without rejecting obvious facts?
- Does the question invite examples, criteria, or trade-offs?
- Are mention candidates relevant and justified?
### 8. Publishing completeness
**5:** Article, source ledger, topic map, cover plan, SEO settings, commentary, disclosure, mentions, and measurement plan are ready and internally consistent.
**4:** Core package is complete with a small operational omission.
**3:** Article is usable but publishing settings or handoff details need work.
**2:** Missing important SEO, visual, disclosure, or packaging elements.
**1:** Public copy and internal notes are mixed, or the package cannot be handed off safely.
Check:
- SEO title is at most 60 characters.
- SEO description targets 140-160 characters.
- Cover direction and alt text are present.
- Feed commentary adds value instead of repeating the headline.
- Hashtags are relevant or omitted.
- Newsletter cadence is realistic when applicable.
- Publication state is accurate.
## Priority Levels
- **Critical:** blocking issue, wrong topic/audience, false claim, fabricated evidence, plagiarism, missing disclosure, or unsafe advice.
- **High:** weak thesis, unsupported material claim, headline mismatch, generic opening, missing application, absent limitation, or broken publishing separation.
- **Medium:** thin example, broad audience language, dense paragraph, weak transition, unclear mention reason, or incomplete measurement note.
- **Low:** small wording, formatting, alt-text, hashtag, or metadata refinement.
## Revision Order
1. Fix scope, audience, thesis, and factual blockers.
2. Resolve unsupported claims and disclosure.
3. Add missing business stakeholders, economics, implementation, measurement, risks, scenarios, and next actions.
4. Add missing decision criteria, application, and limitations.
5. Improve opening, section order, headings, and conclusion.
6. Remove promotional or mechanical language.
7. Run shared humanization and the LinkedIn-specific final 去 AI 化编辑 without changing facts or authority boundaries.
8. Recheck first-person provenance, facts, links, disclosure, SEO fields, and certainty.
9. Validate SEO settings and publishing pack.
10. Run the deterministic audit and recheck changed passages.
## Deterministic vs Editorial Judgment
The script can check:
- H1/H2 counts;
- editorial word-count and paragraph-density warnings;
- primary topic phrase placement;
- image references;
- presence of a LinkedIn Publishing Pack;
- SEO title and description lengths when written in the expected labels;
- presence of a discussion question signal;
- hashtag count in the publishing pack;
- clusters of generic openings, hype, or unverified first-person claims.
- clusters of templated, ceremonial, or synthetic-sounding LinkedIn style signals;
- presence of several common business-depth signals such as economics, ownership, implementation, risk, measurement, and scenarios.
The script cannot determine:
- whether the topic is actually current;
- whether evidence supports a claim;
- whether an idea is original;
- whether the author's experience is real or authorized;
- whether a recommendation is good;
- whether a discussion question will receive comments;
- whether an article was published or performed well.
- whether prose was written by a human or AI, or whether any detector will classify it a particular way.
Never turn a clean script result into a promise of quality or reach.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |