⏳ This skill is pending AI review.
Scores will appear once the review pipeline completes.
linkedin-post
Draft a daily LinkedIn post for S M Moshiur Rahman about business development, grounded in a real problem he hit that day and real numbers from OmniRank. Use when he says "linkedin post", "today's post", "draft a post", or describes friction he wants written up.
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
// README
OmniRank
One page. Every engine.
Score a site's SEO, AEO and GEO signals against its real HTML, then generate the machine-readable ground truth AI engines can cite.
OmniRank scores a website's live HTML against the concrete rules that search engines,
answer engines and generative AI systems use to crawl, quote and cite content, then
generates llms.txt, llms-full.txt and facts.json so the page becomes machine-readable
ground truth an AI engine can cite directly.

Real terminal output from omnirank audit https://example.com — nothing staged, nothing cropped.
SEO, AEO and GEO read the same HTML
Three audiences parse one document. A crawlable, well-marked-up, machine-readable page wins all three at once — that is the entire bet OmniRank makes, and why it audits one artifact (your rendered HTML) instead of three separate tools.
| Layer | Audience | What it wants | What OmniRank does |
|---|---|---|---|
| SEO | Googlebot, Bingbot | Crawlable, canonical, correctly sized metadata, valid structured data | Checks h1, canonical, title/description length, OpenGraph, JSON-LD |
| AEO | AI Overviews, Copilot, voice assistants | A short, liftable, factual answer near the top of the page | Checks for a 40–60 word AnswerBlock, an FAQ section, and speakable markup |
| GEO | ChatGPT, Claude, Perplexity, Gemini | Machine-ingestible ground truth plus explicit permission to cite | Generates and verifies llms.txt, llms-full.txt, facts.json with a citation licence |
The table above is not aspirational copy — it is also, deliberately, how this README is
written. It has one <h1>, an AnswerBlock-shaped opening paragraph, an FAQ section, and
descriptive headings. If OmniRank asked your page to do these things, it does them too.
Audit any site in one command
OmniRank is not published to PyPI, npm or a container registry — install it from source
into a virtual environment. Homebrew and other PEP 668-managed Python installs refuse a
bare pip install, so activate a venv first:
git clone https://github.com/bemoshiur/OmniRank.git && cd OmniRank
python3 -m venv .venv && source .venv/bin/activate && make install
python3 -m omnirank.cli audit https://example.com
Real output, captured against https://example.com:
OmniRank 0.2.0 — https://example.com
overall 76/100 aeo 80 geo 60 perf 100 seo 67
1 URLs checked, 10 findings
[FAIL] seo.canonical.missing https://example.com/
observed: no rel=canonical
fix: Add <link rel="canonical" href="https://example.com/"> to <head>.
[FAIL] seo.description.missing https://example.com/
observed: no meta description
fix: Add a meta description summarising the page.
[WARN] seo.og.missing https://example.com/
observed: missing og:title, og:image
fix: Add the missing OpenGraph tags so social unfurls render.
[FAIL] aeo.answer-block.missing https://example.com/
observed: no element matching '.answer-block'
fix: Add <div class="answer-block" data-speakable> with a 40-60 word plain-prose answer.
[FAIL] aeo.faq.too-few https://example.com/
observed: 0 FAQ pairs
fix: Add FAQs as semantic <dl>/<dt>/<dd> or <details>, mirrored by FAQPage JSON-LD.
[FAIL] seo.schema.absent https://example.com/
observed: no application/ld+json blocks
fix: Emit JSON-LD describing this page and cross-reference the site organisation by stable @id.
[FAIL] geo.llms.missing https://example.com/llms.txt
observed: HTTP 404 at llms.txt
fix: Generate llms.txt at build time and serve it as a static file.
[FAIL] geo.llms-full.missing https://example.com/llms-full.txt
observed: HTTP 404 at llms-full.txt
fix: Generate llms-full.txt at build time and serve it as a static file.
[FAIL] geo.facts.missing https://example.com/facts.json
observed: HTTP 404 at facts.json
fix: Generate facts.json at build time and serve it as a static file.
[FAIL] geo.ai-allowlist.missing https://example.com/robots.txt
observed: HTTP 404 at /robots.txt
fix: Publish a robots.txt that explicitly allows AI crawlers.
report: .omnirank/reports/2026-08-03-audit.json
example.com ships nothing but a static placeholder page, so this is close to a worst
case — real sites usually clear a handful of these on the first pass. Every finding
carries observed, expected, and fix — enough to act on without opening the code.
For a configured repo, copy the starter config and gate a build on specific gates:
cp templates/omnirank.config.example.json omnirank.config.json
python3 -m omnirank.cli audit --config omnirank.config.json --fail-on h1 canonical schema
python3 -m omnirank.cli geo --config omnirank.config.json --out public
Exit codes: 0 clean · 1 a --fail-on gate failed · 2 usage or config error. The
non-zero exit is what makes it a CI gate, not just a report — see
Use OmniRank as a CI gate below.
What ships in v0.4.0
| Skill | Status | What it does |
|---|---|---|
audit | Shipped | Scores SEO, AEO, GEO, perf, security and structured-data gates against a site's real HTML — including a cross-URL pass over the whole crawled set — and reports observed / expected / fix for every gap |
geo-artifacts | Shipped | Generates llms.txt, llms-full.txt and facts.json, each with an explicit citation licence |
fix (preview) | Shipped | Resolves a finding's URL to the source file that owns it and prints the unified diff it would apply for the four mechanical findings. Writes nothing — there is no --write flag |
Still exactly two skills. v0.4.0 adds 18 new finding ids across four groups: a new
security layer (HSTS, nosniff, CSP absence, Referrer-Policy, mixed content, and a
missing HTTP→HTTPS redirect); indexability contradictions provable from a site's own
declarations (a sitemap URL its own robots.txt disallows, a canonical pointing at a
noindexed/redirecting/missing page, a hreflang alternate that is itself noindexed);
Google's required rich-result properties for eight structured-data types; and on-page
accessibility overlap (missing alt, skipped heading levels, empty or generic link
text, missing lang). It also normalises the score model so a layer's budget scales
with its own gate count instead of a flat constant — scores from 0.3.0 and 0.4.0 are
not comparable. omnirank fix still writes nothing to your project; file
modification ships once the locator is proven against real repositories and the write
guarantees in docs/research/2026-08-04-automation-architecture.md §2.5 are
implemented and tested. Everything under Roadmap below is not present in v0.4.0.
Roadmap
| Skill | Target | What it will do |
|---|---|---|
aeo-onpage | v0.2 | Emit JSON-LD by entity type; draft AnswerBlocks and FAQs; Next.js codegen |
indexing | v0.3 | IndexNow, GSC URL Inspection, Bing Submit, Wayback, hash-based freshness |
offsite-entity | v0.4 | sameAs gap analysis, peer mention-ga |
// HOW IT'S BUILT
// FILES
2 files shipped with this skill. Inline text is the catalog copy. The upstream link is the source repository.
.claude/skills/linkedin-post/SKILL.md4.2 KBd0d91266ea39upstream ↗
--- name: linkedin-post description: Draft a daily LinkedIn post for S M Moshiur Rahman about business development, grounded in a real problem he hit that day and real numbers from OmniRank. Use when he says "linkedin post", "today's post", "draft a post", or describes friction he wants written up. --- # Daily LinkedIn post Draft one post in Moshiur's voice. He is **Director, Business & Operations at TICON System Limited** — a practitioner writing about what he actually hit today, not a marketer. ## The one rule **A post needs a real thing that happened.** If he hasn't given you one, ask for it before drafting. Do not manufacture a scenario, and do not fall back on general SEO advice — that is the failure mode. Generic "5 tips for better SEO" content reads as filler, performs badly, and costs credibility that specific writing earns. Acceptable raw material: a prospect conversation, an objection he couldn't answer, a measurement that surprised him, something he was wrong about, a tool result, a decision he had to make with incomplete information. ## Structure that works 1. **A hook that is a fact, not a claim.** Ideally a number or a short admission. "I ran our own website through a tool I built. 57 pages. Not one had an answer an AI could quote." Never open with "In today's digital landscape". 2. **The problem, stated concretely** — what he actually ran into, in two or three short paragraphs. Short lines. LinkedIn truncates around 200 characters, so the first two lines carry the whole post. 3. **Evidence.** Real numbers, indented as a small block. See below for pulling them. 4. **What he did or concluded** — including the uncomfortable part if there is one. 5. **One specific question.** Not "thoughts?" — something a reader can answer in a sentence from their own experience. 6. **Link only if it earns its place.** github.com/bemoshiur/OmniRank when relevant. 7. **4–7 hashtags**, lowercase-natural, no keyword stuffing. Target 1,300–2,000 characters. Longer gets collapsed behind "see more". ## Getting real numbers Never invent a statistic. When a post would benefit from data, measure it: ```bash cd "<repo>" && .venv/bin/python -m omnirank.cli audit https://<site> --out /tmp/post.json ``` Then read `/tmp/post.json` for exact per-layer scores and finding counts. Quote what the tool actually printed. If a number can't be measured, cut the claim rather than soften it. Research already in the repo that can be cited honestly — with its caveats intact — lives in `docs/research/`. The brand-mention and AI-citation correlations there are **observational, mostly English-language B2B, and directional**. Say so if you use them. ## Voice - First person, past tense, specific. - Admit the unflattering part. A Director saying his own site scored 33/100 is more credible than any polished claim, because nobody fakes that direction. - No "game-changer", "leverage", "unlock", "in today's world", "the truth is". - No em-dash-heavy rhetorical flourishes. Plain sentences. - Never claim a result the work hasn't produced. No client names without permission. ## Angles not yet used Track these so the daily cadence doesn't repeat itself. Cross one off when used. - [x] Ran my own company's site through my own tool (used 2026-08-05) - [ ] A prospect asked a question I couldn't answer with data, so I built the measurement - [ ] The thing I believed about SEO that turned out to be wrong when measured - [ ] Why our analytics under-report AI traffic, and by roughly how much - [ ] A bug in my own tool that would have given a client bad advice — and what I changed - [ ] What a Bangladeshi B2B buyer actually does before contacting a vendor - [ ] The difference between ranking and being cited, with numbers - [ ] Why I removed a feature instead of shipping it - [ ] What "AI-ready content" means concretely, at the markup level - [ ] A competitor comparison I ran that didn't flatter us ## Before handing it over State plainly that **you cannot publish to LinkedIn** — there is no integration. Write the post to a file, give him the path, and let him paste it. Flag anything in the draft that exposes the company publicly (a poor score, an internal mistake) so he can decide, rather than deciding for him.README.mdupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
Plain-text copy, no login: /skill-bundles/bemoshiur-omnirank-claude-skills-linkedin-post.md