Agent skills are tremendously useful. They act like modular upgrades, giving AI the precise capabilities it needs to handle specialized tasks. Great example from @SeoRobertson π
We've built around 200 skills at our agency.
None of them hold client-specific guidance. Wherever a step needs it, the skill tells the model to find that client's canonical doc, and to create it if it's missing.
We call them skill hooks.
If I understand this correctly, @dejanseo, by turning off Google Search the moment you/I attach a file, Gemini is essentially locking the door to the room: while the AI is holding the private folder, it is not allowed to reach out to the public internet at allβneither to type search phrases out, nor to bring random web pages in. Have you gained more insights?
Google disabled Google search in my @GeminiApp because I added an attachment, it said:
"Uploading a document to Gemini Chat now locks the session to those files to prevent confidential document content from leaking into external search queries and to stop outside web results from contaminating your file data."
Somebody thought this was a great idea and it passed all the checks and went to production.
WOOOOOOOOOOOOOOOOOW.
After spending the afternoon building, I'm fairly confident that Gemini 4 Argon plus my OKF driven knowledge brain can do a better job than my biological brain alone when it comes to data analysis and consulting.
They weren't kidding when they said it was good at knowledge work.
I'm half excited and half terrified. We are nearing the point where AI will be as smart as any human on any cognitive task.
Confused by Google's overlapping "skills" terminology? π€―
From Gemini productivity features to executable SKILL.md files, bridging human expertise with machine-readable data is vital for agentic SEO.
Google rewards people-first content that fulfills human needs, regardless of how it is produced. To perform in AI search experiences that synthesize complex queries, structured data must explicitly match visible copy. By speaking the deterministic language of machines, we protect and elevate the human experience.
Learn how to structure your Knowledge Graph for the era of AI search β¬οΈ
https://t.co/lESFNvWR1z
#SEO #AgenticAI #KnowledgeGraph #Gemini #GenerativeSearch
Exciting news: Gemini Agent is an autonomous digital co-worker useful in:
- workflows
- cloud databases
- 3rd party business apps
@Marie_Haynes explains this beautifully
Oooohβ¦Google announced Gemini Agent - meant to help small businesses agentically. This is big news!
It works inside Gmail, Drive, Docs, Slides, Sheets, Chat and Calendar carrying the same memory, skills and controls it has everywhere else.
It also has skills for data analysis, industry specific tools, connectors (like DropBox, Asana, GitHub, Notion, Slack) and knowledge.
It runs in the cloud and can be accessed from any device.
It can create subagents to go get stuff done for you in parallel and run for hours or days.
It can use Gemini OR Claude and other models are coming soon including open models.
https://t.co/BdfXjDBlU7
https://t.co/TZ0ihPNyPh
Our Search Bar Pro meeting is starting soon so we will definitely go over this in more detail.
Google announced new YouTube playlists for learning Search fundamentals and Technical SEO. This is such a good idea with resources you can have new SEO staff go through. Or really for anyone who wants to learn SEO.
You'll learn how Google crawls, indexes and serves content, how to optimize for Search and how to use GSC and Google Trends.
These videos are not new - it's just the playlist curation that is.
The playlist for Search fundamentals:
https://t.co/XtrCAS084e
The playlist for Technical Fundamentals:
https://t.co/FRrVzGhuWv
The blog post about these new playlists:
https://t.co/vMkTrALH4g
It would be impossible to write this much without in-depth topic expertise. Congratulations to @rustybrick for an astounding accomplishment. You have a dedicated following that deeply appreciates your authorship. π