agentsclimarketplace

Real estate youtube aeo

Skill ttventuri/real-estate-youtube-aeo

Convert real estate video, podcast, webinar, market-update, neighborhood, property, buyer, seller, mortgage, or relocation transcripts into accurate, paste-ready YouTube publishing packages designed for SEO, answer-engine optimization (AEO), generative-engine optimization (GEO), local relevance, and machine understanding. Use when a real estate agent, team, brokerage, lender, title professional, or real estate media creator supplies a transcript and wants an optimized title, description, chapters, transcript-grounded Q&A, tags, hashtags, pinned comment, playlist recommendation, thumbnail concept, compliance review, or future recording notes. Also use to revise existing YouTube metadata from a transcript. Do not use for channel-wide audits, performance diagnosis, competitor research, or keyword-volume research unless the user separately requests those tasks.From its SKILL.md

Install
npx -y skills add ttventuri/real-estate-youtube-aeo

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

8.0 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it

Real Estate YouTube AEO Optimizer

Turn a real estate transcript into a clear, accurate YouTube publishing package. Optimize for durable understanding and search relevance without promising rankings, indexing, traffic, citations, AI Overview inclusion, or LLM references.

Portability and neutrality

  • Follow the open Agent Skills structure: this SKILL.md is the entry point and the references/ files are supporting context.
  • Keep the workflow independent of any specific AI assistant, vendor, model, operating system, or tool name.
  • Do not assume or hardcode any agent, team, brokerage, lender, title company, channel, speaker, market, city, state, country, service area, link, playlist, disclosure, or brand voice.
  • Treat brand and location as optional user-supplied context. Omit them when absent or irrelevant.
  • Use capabilities available in the current environment. If a requested capability is unavailable, complete the remaining package and state the limitation briefly.
  • Do not require web access, external APIs, image generation, or proprietary integrations for the core transcript-to-metadata workflow.

Inputs

Require only a transcript. Accept these optional inputs:

  • optional creator, organization, or channel profile
  • current or proposed title
  • recording date or market-data period
  • speaker, guest, and organization
  • episode number or series name
  • target audience, geographic market, or primary question
  • website, contact, listing, market-report, or social links
  • existing playlists and title archive
  • required brokerage, licensing, or advertising disclosures
  • thumbnail direction and accessible identity-reference images

Do not block on missing optional inputs. Read references/channel-profile.md to normalize supplied brand and market information. Omit unavailable links or boilerplate rather than inventing them.

Load references

Always read:

  • references/optimization-guidance.md
  • references/output-format.md
  • references/compliance.md

Read references/channel-profile.md when brand, market, links, playlists, speakers, or disclosures are supplied or need to be separated from transcript facts.

Read references/thumbnail-style.md before proposing or building a thumbnail.

Workflow

  1. Read the complete transcript before drafting.
  2. Classify the episode type and freshness: evergreen, dated/current, recurring update, property-specific, or event-specific.
  3. Identify one primary search intent, one core question, and the clearest direct answer supported by the transcript.
  4. Extract verified speakers, organizations, locations, dates, statistics, sources, property facts, and real estate concepts. Do not infer missing facts.
  5. Separate transcript facts from channel-profile facts. Use each only where relevant.
  6. Determine whether the intended search phrase, location, and direct answer were spoken clearly. Flag missing language in future recording notes.
  7. Check the transcript and proposed metadata for fair-housing, advertising, financial, legal, statistical, and brand-safety risks.
  8. If a title archive is supplied, check exact duplication, near duplication, repetitive phrasing, and recently reused angles.
  9. Draft a specific title that leads with the subject a viewer would search. Include a city, region, or market only when it is central and supported.
  10. Draft a description whose first sentence states the topic, any relevant location, and the main takeaway. Put unique episode content before channel boilerplate.
  11. Add three or four concise Q: and A: pairs grounded in the dominant topic.
  12. Add chapters only when timestamps support them. Start at 00:00, keep them ascending, and use searchable labels. Mark approximate or edit-shifted timings for human review.
  13. Generate restrained tags and three to five relevant hashtags. Treat tags as secondary metadata.
  14. Recommend an existing playlist only when the supplied profile or archive confirms it. Otherwise recommend a playlist category without inventing a URL.
  15. Propose one thumbnail concept, but do not generate the image yet.
  16. Run grounding, compliance, statistical-integrity, and formatting checks before returning the final package.

Grounding rules

  • Trust the transcript for episode-specific facts.
  • Use the channel profile only for verified identity, service area, links, playlists, boilerplate, and required disclosures.
  • Do not add local keywords merely because the creator works in a particular market; the video must meaningfully support the location.
  • Never invent statistics, sources, guests, locations, conclusions, links, credentials, property facts, or brokerage relationships.
  • Preserve material nuance. Do not turn a conditional observation into a universal claim.
  • Attribute statistics and identify their time period, geography, source, and property type when supplied.
  • Distinguish average from median, price mix from appreciation, list price from sale price, and local results from national results.
  • Do not imply that metadata alone makes a video rank, appear in AI answers, or become cited by an LLM.
  • Avoid keyword stuffing and lists of cities or neighborhoods the video does not meaningfully discuss.
  • Prefer full place names over unexplained abbreviations.
  • Use REALTOR® only when the user identifies the person or organization as an NAR member and wants that term; otherwise use real estate agent, broker, or the transcript's accurate role.

Thumbnail approval gate

Provide the concept first and end with the approval question specified in references/output-format.md.

If the user approves or clearly asks to build:

  1. Confirm that an identity-reference image is accessible in the current session or through a user-supplied file path.
  2. If missing, ask the user to provide it and stop before image generation.
  3. If present, use it as the direct identity reference and follow references/thumbnail-style.md.
  4. If the environment provides image-generation or image-editing capability, use it. Otherwise return the approved production prompt without pretending the image was built.
  5. Revise if the person is not clearly recognizable.

Optional website amplification

Do not include a companion webpage by default. If the user requests stronger Google or LLM discoverability, a blog, or a video page, offer or produce a separate package with:

  • semantic lede and key takeaways
  • transcript-grounded Q&A
  • suggested slug and meta description
  • internal-link suggestions based on supplied pages
  • VideoObject inputs
  • visible transcript or transcript-summary guidance
  • factual source and date notes

Keep this separate from the paste-ready YouTube description.

Quality bar

Before responding, confirm:

  • the title matches the dominant transcript topic
  • the first description sentence provides a direct semantic summary
  • every Q&A answer is supported by the transcript
  • chapters are usable or marked for review
  • names, locations, credentials, and brokerage relationships are verified
  • claims and statistics retain their qualifications
  • compliance risks are flagged without adding steering language
  • the final paste-ready block contains no internal analysis or unsupported placeholders

What ships with it: 6 files

13.7 KB alongside SKILL.md

agents/

Keep looking

Skills are one crate of 326,144. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.