Hey folks, 👋
Today’s publication is JUICY.
📌 In a nutshell
Claude is now in Slack → DM it, @mention it in threads, and your Skills work too.
Adobe apps in ChatGPT → Photoshop, Express, and Acrobat (and all free!). Merge PDFs, edit images, create designs through conversation.
Manus Browser Operator → BYO browser. Install their extension on Chrome, and their AI agent complete browser based tasks for you.
Google Pomelli added animation → Cool tool with potential, but rarely produces usable assets. Ignore the "replace your marketing team" hype.
2 Claude Skills you can steal → An image-gen prompt writing skill trained on Google's official docs, and a JSON prompting skill for reproducible outputs. I used these to create a custom 3D icon set.
Use Adobe tools in ChatGPT (for free)
Adobe just brought three of its flagship apps into ChatGPT - Photoshop, Adobe Express, and Acrobat. All free. All available now.
You can now tell ChatGPT things like "blur the background of this photo" or "merge these PDFs" and Adobe's tools handle it right inside the chat!
Can confirm this is a pretty cool, and very useful update!
Photoshop lets you adjust brightness, contrast, and exposure, remove or blur backgrounds, and apply creative effects. You get sliders to fine-tune results without leaving the conversation.
Adobe Express pulls from its template library to generate designs for invitations, social posts, and graphics. You can iterate on edits and even animate designs directly in chat.
Acrobat is the sleeper hit here. I tested it, and it works great. You can merge multiple PDFs, extract text or tables from documents, compress files, edit text in existing PDFs, and redact sensitive info - all through conversation.

Claude is now in Slack
Anthropic just shipped a proper Slack integration, and it's more useful than I expected.
You can now DM Claude directly in Slack, @mention it in any thread, or pull it up from the AI assistant panel. It works like regular Claude, but it lives where your conversations already happen.
I tested it and your Skills work inside Slack!! If you've built a brand voice skill, a content repurposing workflow, or a hook writing skill, Claude can now apply them without you leaving Slack.

Use cases I'm excited about:
Meeting prep from Slack context: Have Claude pull together talking points from recent channel discussions before a call.
Turn messy threads into docs: Ask Claude to formalise a 40-message discussion into an SOP, brief, or summary.
Summarise what you missed: Catch up on long threads without reading every message.
Research in the flow of conversation: Quick answers without context-switching to another app.
Currently it’s missing: Research mode, file creation, or extended thinking yet. Also no Projects access (each conversation is fresh). And it requires a paid Slack plan
I can see this evolving into something more powerful. You can install here.
For the developers: Anthropic also launched Claude Code in Slack as a research preview. You can now @mention Claude in a thread discussing a bug or feature request, and it'll spin up a full coding session - analysing the conversation, picking the right repo, posting progress updates, and opening a pull request. All without leaving Slack. If you're already using Claude Code, this looks NUTS.
Manus just turned any browser into an AI browser
The main appeal with AI browsers like Atlas and Comet is AI agents that can do browser based tasks for you like; find a working discount code, analyse a YouTube channel, scrape competitor data, send LinkedIn connections to a list. But switching browsers is a pain, and I'm used to Chrome.
Manus took a different approach: BYO browser. They're a Chinese company, and if you know anything about Chinese restaurants, they pioneered BYO alcohol. Bring your own bottle, we'll handle the food. Same energy here. Keep your browser, they supply the agent.
Their new Browser Operator extension bolts onto your favourite browser, like Chrome or Edge. Install it, connect your account, and Manus can do tasks.
The clever bit: because it runs inside your actual browser session (your IP, your cookies, your logins), it bypasses CAPTCHAs and auth walls that trip up cloud-based agents.
One small note: this does require granting broad permissions. It's not storing your passwords, but it is operating inside your authenticated sessions. Worth being aware of.
Animation added to Google Pomelli
Pomelli is a free AI marketing tool that helps small- to medium-sized businesses create on-brand marketing campaigns.
It works by scanning your website to build a “Business DNA” profile (capturing your brand’s tone, fonts, colour palette, imagery and style), then generating tailored campaign ideas and editable creative assets (posts, ads, visuals) that match your brand identity, ready to be used across social media or other marketing channels.
From my personal experience… Pomelli is a cool tool with lots of potential, but it’s a little bit gimmicky and rarely produces usable assets - and i’m sick of seeing videos about it saying “this tool will replace your marketing team”.
💡 Builder’s notes
CLAUDE OPUS 4.5 IS HANDS DOWN THE BEST AI MODEL I’VE EVER USED. It’s great at writing, great at using tools, great at coding. I had to upgrade to the $200 per month plan for max usage 😂. I still think Gemini 3 Pro sucks at daily tasks - change my mind.
I built 2 cool skills in Claude this week (if you need a refresh on skills watch this):
An image-gen prompt writing skill
I fed Claude all of Googles official documentation on how to write prompts for Nanobanana (Googles image generation AI). The same prompt writing principles pretty much apply to all of the other AI image models too. So now when I can ask Claude “write me a prompt to generate an image of a male fitness influencer holding this product”, Claude can write me outstanding prompts, that I can then feed that prompt into an image model to get results like this…

A JSON image prompting skill
A skill for creating structured JSON prompts for AI image generators (NanoBanana, Kling, Flux, etc.). Unlike the text-based image prompting skill which outputs plain prose prompts, this outputs machine-readable JSON that's easier to edit, iterate on, and reproduce consistently. JSON is better when you have a hyper-specific style you want to recreate consistently I used this skill to create a cool custom 3D icon set. You can watch my mini tutorial here.

If you want access to these Skills + plenty more for FREE…
Just share this newsletter with 1 friend (who’s into AI), and it’s all yours! You currently have {{rp_num_referrals}} referrals, only {{rp_num_referrals_until_next_milestone}} away.
or send them this link: {{rp_refer_url}}
🥣 Brain food
50+ Claude's use cases → Get inspired by what you can do with Claude. Practical examples across research, writing, coding, analysis, and everyday tasks.
Immaculate use of free will → lol (✅LOL verified at 13:23:36 11/12/2025)
Steve Jobs' lifechanging speech → Listen while you're eating lunch today. Seriously inspiring. "You can't connect the dots looking forward; you can only connect them looking backwards."
Have a splendid rest of your week, and I’ll catch you guys next Thursday!! Rumour has it ChatGPT 5.2 will be released.
To your success,













