Loading video...

Video Failed to Load

Go Home

SOMEONE BUILT A CLAUDE CODE PLUGIN THAT REFUSES TO WORK UNTIL YOU DO PUSH-UPS OR SQUATS its called workout gate and it is exactly as cursed as it sounds when the hook fires, it blocks your prompt, opens your webcam, and makes you exercise to keep coding > it...

om_patel5's profile picture

Om Patel

27,170 subscribers

671,976 views • 5 months ago •via X (Twitter)

Health & WellnessEducationScience & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

You can’t control what happens when it’s time to compete, but you can control how prepared you are for it. Get your work in. Get your reps. Be intentional with your time. No one can take your preparation away from you. Gianna Selby working cuts and relays here with a fungo off the bat. Reading the flight and speed of the ball and working on timing the catch: right foot in the air as the ball hits the glove so we can get it back in the ground and work right/left into the throw.
0:14

You can’t control what happens when it’s time to compete, but you can control how prepared you are for it. Get your work in. Get your reps. Be intentional with your time. No one can take your preparation away from you. Gianna Selby working cuts and relays here with a fungo off the bat. Reading the flight and speed of the ball and working on timing the catch: right foot in the air as the ball hits the glove so we can get it back in the ground and work right/left into the throw.

Coach Steinman

35,715 views • 1 month ago

We've updated /fork in Claude Code /fork now runs a background agent with your exact context (system prompt, tools, history, model) and prompt cache. The result gets returned to your session. /branch (the old /fork) still copies the transcript to a new session you drive.
0:16

We've updated /fork in Claude Code /fork now runs a background agent with your exact context (system prompt, tools, history, model) and prompt cache. The result gets returned to your session. /branch (the old /fork) still copies the transcript to a new session you drive.

ClaudeDevs

381,886 views • 5 months ago

Pika just launched their 4K-VFX skill through MCP and I edited this entire video on Claude Code. When you want to replicate the viral video and realized you can just shoot it in your office parking lot and edit it in Claude Code... 🤤 Get my prompt in the thread:
0:14

Pika just launched their 4K-VFX skill through MCP and I edited this entire video on Claude Code. When you want to replicate the viral video and realized you can just shoot it in your office parking lot and edit it in Claude Code... 🤤 Get my prompt in the thread:

Jessie_Ma

217,032 views • 4 months ago

You can now run Grok inside Claude Code and Codex, using your existing subscription. it's called grok-plugin. Grok’s genuinely good right now, but if your setup lives in Claude Code and Codex, switching to use it is a pain. this solves that exact issue: > Grok in Claude Code and Codex > directly auths through X Premium or Supergrok, no api key > your workflow and subagents stay untouched free and open source. the low-friction way to test Grok in the tools you already use.
0:16

You can now run Grok inside Claude Code and Codex, using your existing subscription. it's called grok-plugin. Grok’s genuinely good right now, but if your setup lives in Claude Code and Codex, switching to use it is a pain. this solves that exact issue: > Grok in Claude Code and Codex > directly auths through X Premium or Supergrok, no api key > your workflow and subagents stay untouched free and open source. the low-friction way to test Grok in the tools you already use.

Alvaro Cintas

26,018 views • 2 months ago

Just made this with Seedance. Full workflow, start to finish: - Grab a reference image — TikTok, Instagram, or Pinterest - Upload it to Claude, write a JSON prompt, adjust as you like - Generate the image - Upload it to Seedance 2.5 with your full video prompt Done All inside MakeUGC, for $1. I made a full guide breaking down the exact JSON prompt structure that makes this work. Comment "JSON" and I'll send it over
0:15

Just made this with Seedance. Full workflow, start to finish: - Grab a reference image — TikTok, Instagram, or Pinterest - Upload it to Claude, write a JSON prompt, adjust as you like - Generate the image - Upload it to Seedance 2.5 with your full video prompt Done All inside MakeUGC, for $1. I made a full guide breaking down the exact JSON prompt structure that makes this work. Comment "JSON" and I'll send it over

Cas.Fyn

17,901 views • 2 months ago

🚨 Anthropic just made Claude Code remote-controlled. Meet Channels. You can now: → Connect a persistent local Claude session directly to a Discord or Telegram Bot → Message tasks to your bot on the go → Come back to finished code 🔥 Here’s how to set it up: 1️⃣ Create a bot (via Telegram’s BotFather or the Discord Dev Portal) and copy the token 2️⃣ Inside Claude, run /plugin install telegram@claude-plugins-official (or discord) 3️⃣ Configure the token and relaunch your session with the --channels flag 4️⃣ DM your new bot, grab the 6-character code, and pair it to your session! Message the bot to fetch files, read history, and execute code locally while you are on the go. The flexibility is unmatched. Time to spin up your bot 👀
0:17

🚨 Anthropic just made Claude Code remote-controlled. Meet Channels. You can now: → Connect a persistent local Claude session directly to a Discord or Telegram Bot → Message tasks to your bot on the go → Come back to finished code 🔥 Here’s how to set it up: 1️⃣ Create a bot (via Telegram’s BotFather or the Discord Dev Portal) and copy the token 2️⃣ Inside Claude, run /plugin install telegram@claude-plugins-official (or discord) 3️⃣ Configure the token and relaunch your session with the --channels flag 4️⃣ DM your new bot, grab the 6-character code, and pair it to your session! Message the bot to fetch files, read history, and execute code locally while you are on the go. The flexibility is unmatched. Time to spin up your bot 👀

Charly Wargnier

14,611 views • 7 months ago

The new Anthropic prompt improver is incredible. You just need to write your prompt and Claude will generate an optimized prompt instantly. Here’s how to use it:
0:16

The new Anthropic prompt improver is incredible. You just need to write your prompt and Claude will generate an optimized prompt instantly. Here’s how to use it:

Alvaro Cintas

205,988 views • 2 years ago

Adding motion graphics to a video used to take hours. Now all it takes is a prompt. With the ElevenLabs MCP and Claude Motion, you can create videos and add motion graphics without leaving Claude. Read the article below to learn how, and save it for your next project.
0:13

Adding motion graphics to a video used to take hours. Now all it takes is a prompt. With the ElevenLabs MCP and Claude Motion, you can create videos and add motion graphics without leaving Claude. Read the article below to learn how, and save it for your next project.

ElevenCreative

132,732 views • 1 month ago

New in Claude Code: your sessions can now message each other. Instead of having to re-explain yourself in another session, you can now tell Claude to do it. It sends a summary (not your history or files), and the other session picks it up mid-task.
0:16

New in Claude Code: your sessions can now message each other. Instead of having to re-explain yourself in another session, you can now tell Claude to do it. It sends a summary (not your history or files), and the other session picks it up mid-task.

ClaudeDevs

7,989,508 views • 3 months ago

Give Opus 5.5 𝗮 𝗰𝗹𝗶𝗰𝗸𝗮𝗯𝗹𝗲 𝗺𝗼𝘁𝗶𝗼𝗻 𝗱𝗶𝗰𝘁𝗶𝗼𝗻𝗮𝗿𝘆: 𝟭𝟮𝟬 𝗺𝗼𝘃𝗶𝗻𝗴 𝘀𝗮𝗺𝗽𝗹𝗲𝘀, each labelled with its name and a ready-made prompt. Click the ones you like, and Claude Code builds them into your video or site. Five packs: motion graphics, camera moves and transitions, UI components, landing page and web effects, and photo looks and lighting. Four steps: 1. Paste the line at the end into Claude Code to install the skill 2. Tell it "Open the motion-prompt library" 3. Tick the samples you like in the gallery and press "Send to AI" 4. Back in Claude Code, say "Edit the opening of this video like this" It gets each sample's motion spec and reference code, and it works with Remotion, CSS or Canvas. Without the skill, press "Copy picks" and paste them in. Paste this into Claude Code and it installs itself 👇 "Read and install the motion-prompt skill in this environment. When it is installed, tell me how to use it in three lines."
0:19

Give Opus 5.5 𝗮 𝗰𝗹𝗶𝗰𝗸𝗮𝗯𝗹𝗲 𝗺𝗼𝘁𝗶𝗼𝗻 𝗱𝗶𝗰𝘁𝗶𝗼𝗻𝗮𝗿𝘆: 𝟭𝟮𝟬 𝗺𝗼𝘃𝗶𝗻𝗴 𝘀𝗮𝗺𝗽𝗹𝗲𝘀, each labelled with its name and a ready-made prompt. Click the ones you like, and Claude Code builds them into your video or site. Five packs: motion graphics, camera moves and transitions, UI components, landing page and web effects, and photo looks and lighting. Four steps: 1. Paste the line at the end into Claude Code to install the skill 2. Tell it "Open the motion-prompt library" 3. Tick the samples you like in the gallery and press "Send to AI" 4. Back in Claude Code, say "Edit the opening of this video like this" It gets each sample's motion spec and reference code, and it works with Remotion, CSS or Canvas. Without the skill, press "Copy picks" and paste them in. Paste this into Claude Code and it installs itself 👇 "Read and install the motion-prompt skill in this environment. When it is installed, tell me how to use it in three lines."

Vox

64,536 views • 27 days ago

Someone leaked the system prompts of every major AI, including Fable 5. That means you can basically run Fable at Opus prices Every AI model runs on a hidden system prompt: the secret instructions that shape how it behaves, its tone, its formatting, what it will and won't do. The personality and rules behind the model. Companies never show you these. Now they're public. A GitHub repo (asgeirtj/system_prompts_leaks) collects the leaked system prompts behind Claude, ChatGPT, Gemini, Grok, Copilot and more. Fable 5's is in there, plus a diff showing exactly what Anthropic changed from Opus 4.8. So you can paste Fable 5's prompt into a cheaper model like Opus 4.8, and it'll start behaving like Fable, same style, same habits. You won't get Fable's actual intelligence (that lives in the model), but you get its feel. The real value: it's a masterclass in prompt engineering. The Opus-to-Fable diff shows exactly what Anthropic tuned for their most powerful model. Steal those patterns. How to use it: 1. Open the repo and go to the Anthropic section 2. Open Fable 5's system prompt, or the Opus 4.8 diff 3. Copy it, paste it as the system prompt in your own model or API call 4. Or just read the diff and lift the patterns into your own prompts Bookmark this
0:18

Someone leaked the system prompts of every major AI, including Fable 5. That means you can basically run Fable at Opus prices Every AI model runs on a hidden system prompt: the secret instructions that shape how it behaves, its tone, its formatting, what it will and won't do. The personality and rules behind the model. Companies never show you these. Now they're public. A GitHub repo (asgeirtj/system_prompts_leaks) collects the leaked system prompts behind Claude, ChatGPT, Gemini, Grok, Copilot and more. Fable 5's is in there, plus a diff showing exactly what Anthropic changed from Opus 4.8. So you can paste Fable 5's prompt into a cheaper model like Opus 4.8, and it'll start behaving like Fable, same style, same habits. You won't get Fable's actual intelligence (that lives in the model), but you get its feel. The real value: it's a masterclass in prompt engineering. The Opus-to-Fable diff shows exactly what Anthropic tuned for their most powerful model. Steal those patterns. How to use it: 1. Open the repo and go to the Anthropic section 2. Open Fable 5's system prompt, or the Opus 4.8 diff 3. Copy it, paste it as the system prompt in your own model or API call 4. Or just read the diff and lift the patterns into your own prompts Bookmark this

Yarchi

22,533 views • 4 months ago

Marketers. I’m begging you. Stop prompting Claude 12+ times when it can be done in 1. Most marketers open Claude and start typing whatever comes to mind. Several rounds later, you’re editing mediocre output by hand. My team built a prompt auditing tool that reads your work, grades it, and gets you to 95% in minutes. It’s based on the RCTC framework, which answers the 4 questions Claude actually needs to be effective: Role - WHO is this for? Channel - WHERE is this going? Task - WHAT outcome is the output getting to? Constraints - WHAT bad output sounds like? Then automatically edits your work so you’re more effective in Claude, ChatGPT, or Gemini. Comment “GA” and I’ll send you the tool.
0:13

Marketers. I’m begging you. Stop prompting Claude 12+ times when it can be done in 1. Most marketers open Claude and start typing whatever comes to mind. Several rounds later, you’re editing mediocre output by hand. My team built a prompt auditing tool that reads your work, grades it, and gets you to 95% in minutes. It’s based on the RCTC framework, which answers the 4 questions Claude actually needs to be effective: Role - WHO is this for? Channel - WHERE is this going? Task - WHAT outcome is the output getting to? Constraints - WHAT bad output sounds like? Then automatically edits your work so you’re more effective in Claude, ChatGPT, or Gemini. Comment “GA” and I’ll send you the tool.

Jesse Pujji

10,805 views • 5 months ago

20 days ago, I connected Claude Code to my newly created instagram handle.. I gained 4.3M views and 6500+ followers in less than a month [ i post Ai generated animated stories ] Full workflow: i let claude study my account before i write another reel.. This is the cleanest content workflow i've built on claude. give it your IG first. 4 prompts handle the rest.. niche research, the reel script, the hook, and the daily automation.. the whole loop is basically, give claude your IG → find what's working → write retention-optimized scripts → engineer the hook → automate the daily output.. ▫️ Setup: give claude your instagram open claude code. claude code has a built-in web tool that browses any public URL. or install any agentic browser like Browser Harness or Firecrawl or Comet browser paste this with your handle filled in: "Browse and pull the last 30 reels and posts. Analyze my recurring topics, top-performing hooks, formats, and engagement patterns. Then map out my actual audience and what they consistently respond to." claude reads your profile, pulls every reel down, and now has the context to personalize every prompt below to YOUR account, not a generic niche. if you're on claude desktop, the same works with firecrawl MCP connected. ▫️ Prompt 1 find what actually goes viral in your niche: "Analyze the highest-performing Instagram Reels, TikToks, and Reddit posts in the [niche] niche from the last 30 days. Identify repeating hooks, visual styles, emotional triggers, and content formats that consistently generate high engagement. Then summarize the 5 strongest content angles optimized for AI-generated content and short-form videos." run this after the setup. you get 5 angles backed by what's already working in your niche, cross-checked against what's already working on YOUR account. ▫️ Prompt 2 write a high-retention reel script "Write a short-form Instagram Reel script about [topic] with an aggressive hook in the first 2 seconds. Create immediate curiosity, tension, or controversy to stop scrolling, then deliver a fast and satisfying payoff. Keep it under 30 seconds and optimize the structure for watch time, replays, comments, and shares. Finish with a subtle CTA." the line that matters: "optimize the structure for watch time, replays, comments, and shares." claude writes for the metrics, not just the word count. ▫️ Prompt 3 engineer better hooks "Study the top-performing Reels in [niche] and break down the hook structure, pacing, and emotional triggers used in the first 3 seconds. Then generate 5 new hook variations that are even more curiosity-driven, emotionally charged, and optimized to stop scrolling instantly. Focus on triggers like surprise, fear, ego, urgency, or desire." most reels die in the first 2 seconds. this prompt has claude reverse-engineer what already works, then give you 5 sharper versions to swap in. ▫️ Prompt 4 automate the whole workflow "Build a complete AI-powered content workflow for Instagram in the [niche] niche. The system should identify trending topics daily, generate high-retention scripts, create matching AI visuals, turn them into short-form videos, and generate optimized captions and hashtags. Structure everything as a repeatable workflow designed for consistent daily posting and growth." once the niche and script structure are validated, this turns it into a daily loop. one prompt that handles topic → script → visual → video → caption. these 4 prompts are the building blocks. the setup is what makes them yours. your real value is in the [niche] you plug in. content workflow built in one weekend, daily posting on autopilot from monday.
0:12

20 days ago, I connected Claude Code to my newly created instagram handle.. I gained 4.3M views and 6500+ followers in less than a month [ i post Ai generated animated stories ] Full workflow: i let claude study my account before i write another reel.. This is the cleanest content workflow i've built on claude. give it your IG first. 4 prompts handle the rest.. niche research, the reel script, the hook, and the daily automation.. the whole loop is basically, give claude your IG → find what's working → write retention-optimized scripts → engineer the hook → automate the daily output.. ▫️ Setup: give claude your instagram open claude code. claude code has a built-in web tool that browses any public URL. or install any agentic browser like Browser Harness or Firecrawl or Comet browser paste this with your handle filled in: "Browse and pull the last 30 reels and posts. Analyze my recurring topics, top-performing hooks, formats, and engagement patterns. Then map out my actual audience and what they consistently respond to." claude reads your profile, pulls every reel down, and now has the context to personalize every prompt below to YOUR account, not a generic niche. if you're on claude desktop, the same works with firecrawl MCP connected. ▫️ Prompt 1 find what actually goes viral in your niche: "Analyze the highest-performing Instagram Reels, TikToks, and Reddit posts in the [niche] niche from the last 30 days. Identify repeating hooks, visual styles, emotional triggers, and content formats that consistently generate high engagement. Then summarize the 5 strongest content angles optimized for AI-generated content and short-form videos." run this after the setup. you get 5 angles backed by what's already working in your niche, cross-checked against what's already working on YOUR account. ▫️ Prompt 2 write a high-retention reel script "Write a short-form Instagram Reel script about [topic] with an aggressive hook in the first 2 seconds. Create immediate curiosity, tension, or controversy to stop scrolling, then deliver a fast and satisfying payoff. Keep it under 30 seconds and optimize the structure for watch time, replays, comments, and shares. Finish with a subtle CTA." the line that matters: "optimize the structure for watch time, replays, comments, and shares." claude writes for the metrics, not just the word count. ▫️ Prompt 3 engineer better hooks "Study the top-performing Reels in [niche] and break down the hook structure, pacing, and emotional triggers used in the first 3 seconds. Then generate 5 new hook variations that are even more curiosity-driven, emotionally charged, and optimized to stop scrolling instantly. Focus on triggers like surprise, fear, ego, urgency, or desire." most reels die in the first 2 seconds. this prompt has claude reverse-engineer what already works, then give you 5 sharper versions to swap in. ▫️ Prompt 4 automate the whole workflow "Build a complete AI-powered content workflow for Instagram in the [niche] niche. The system should identify trending topics daily, generate high-retention scripts, create matching AI visuals, turn them into short-form videos, and generate optimized captions and hashtags. Structure everything as a repeatable workflow designed for consistent daily posting and growth." once the niche and script structure are validated, this turns it into a daily loop. one prompt that handles topic → script → visual → video → caption. these 4 prompts are the building blocks. the setup is what makes them yours. your real value is in the [niche] you plug in. content workflow built in one weekend, daily posting on autopilot from monday.

Axel Bitblaze 🪓

203,927 views • 4 months ago

Introducing Time Machine for Claude Code Works in both Claude Desktop and Claude Code CLI /timemachine records your session as a timeline, including prompts, tool calls, edits, Bash commands, and turn endings Pick any point and fork from there, It restores both the conversation and the code state from that moment into a separate Git worktree and branch, while your original session stays untouched Install: npx claude-code-templates@latest --mod productivity/session-time-machine Then run: /reload-plugin /timemachine
0:16

Introducing Time Machine for Claude Code Works in both Claude Desktop and Claude Code CLI /timemachine records your session as a timeline, including prompts, tool calls, edits, Bash commands, and turn endings Pick any point and fork from there, It restores both the conversation and the code state from that moment into a separate Git worktree and branch, while your original session stays untouched Install: npx claude-code-templates@latest --mod productivity/session-time-machine Then run: /reload-plugin /timemachine

Daniel San

24,097 views • 1 month ago

Your Claude Code setup is probably bloated and you don't even know it. Every MCP server you add loads its full tool list into your context window the second your session starts. Stack too many and you're burning context before you type a word. So the move was never "add more." It's "add the right ones." Anthropic dropped an official plugin that does exactly that. It scans your actual codebase and tells you what's worth adding across all five layers: → hooks → skills → MCP servers → subagents → slash commands React project? It suggests Playwright. Auth code? It flags a security reviewer. It's read-only. Touches nothing. Just hands you the blueprint. The plugin isn't the edge. Knowing it exists is. /plugin install claude-code-setup@claude-plugins-official Steal this before everyone does.
0:10

Your Claude Code setup is probably bloated and you don't even know it. Every MCP server you add loads its full tool list into your context window the second your session starts. Stack too many and you're burning context before you type a word. So the move was never "add more." It's "add the right ones." Anthropic dropped an official plugin that does exactly that. It scans your actual codebase and tells you what's worth adding across all five layers: → hooks → skills → MCP servers → subagents → slash commands React project? It suggests Playwright. Auth code? It flags a security reviewer. It's read-only. Touches nothing. Just hands you the blueprint. The plugin isn't the edge. Knowing it exists is. /plugin install claude-code-setup@claude-plugins-official Steal this before everyone does.

Evan Luthra

36,455 views • 4 months ago

Absolutely recommend using Cache Control in Claude Code Probably one of the Mods that will save you the most tokens and help keep your sessions going much longer It adds this bar for Claude’s 5 minute cache and sends notifications to help keep the cache warm Install command: npx claude-code-templates@latest --mod observability/prompt-cache-control
0:14

Absolutely recommend using Cache Control in Claude Code Probably one of the Mods that will save you the most tokens and help keep your sessions going much longer It adds this bar for Claude’s 5 minute cache and sends notifications to help keep the cache warm Install command: npx claude-code-templates@latest --mod observability/prompt-cache-control

Daniel San

235,089 views • 1 month ago

New Feature Live!🚨 Introducing Prompt Assist—a feature designed to streamline your prompt-writing experience with Alchemist AI. Prompt Assist integrates a speed-oriented AI model that analyzes your input as you write, offering contextually relevant suggestions in real-time. Here’s how it works: • Real-Time Suggestions: Prompt Assist dynamically processes your input, predicting next steps from your initial wording. Press Tab to accept suggestions and seamlessly complete your prompt. • Smart Context Analysis: The feature leverages natural language processing models to understand the structure and intent of your prompt, this ensures the suggestions remain precise and useful.
0:19

New Feature Live!🚨 Introducing Prompt Assist—a feature designed to streamline your prompt-writing experience with Alchemist AI. Prompt Assist integrates a speed-oriented AI model that analyzes your input as you write, offering contextually relevant suggestions in real-time. Here’s how it works: • Real-Time Suggestions: Prompt Assist dynamically processes your input, predicting next steps from your initial wording. Press Tab to accept suggestions and seamlessly complete your prompt. • Smart Context Analysis: The feature leverages natural language processing models to understand the structure and intent of your prompt, this ensures the suggestions remain precise and useful.

ALCHEMIST AI 🔮

37,892 views • 1 year ago

🚨 YOUR AI AGENT CAN NOW READ YOUR NOTIFICATIONS AND ACT ON THEM BEFORE YOU EVEN SEE THEM It’s called PhoneUse and it just launched. It gives your agent its own phone in the cloud, a real Android, connected to Claude through an MCP. Time needed: 10 minutes Steps: connect the MCP, tell your agent what to watch for What it handles on its own: → Your flight gets delayed: it sees the push notification and texts whoever is picking you up → Your Uber is on the way: it checks the driver’s status and lets people know when you’ll arrive → An app asks for a 2-step code: it reads the SMS and gets through the verification → An old app with no API: it opens it and taps through it like a person would Until now agents were stuck in the browser. Most of the apps we use every day don’t have an API, so they couldn’t touch them. Now they have a phone. Save it before it gets deleted 🔖 Link in the comments 👇
0:13

🚨 YOUR AI AGENT CAN NOW READ YOUR NOTIFICATIONS AND ACT ON THEM BEFORE YOU EVEN SEE THEM It’s called PhoneUse and it just launched. It gives your agent its own phone in the cloud, a real Android, connected to Claude through an MCP. Time needed: 10 minutes Steps: connect the MCP, tell your agent what to watch for What it handles on its own: → Your flight gets delayed: it sees the push notification and texts whoever is picking you up → Your Uber is on the way: it checks the driver’s status and lets people know when you’ll arrive → An app asks for a 2-step code: it reads the SMS and gets through the verification → An old app with no API: it opens it and taps through it like a person would Until now agents were stuck in the browser. Most of the apps we use every day don’t have an API, so they couldn’t touch them. Now they have a phone. Save it before it gets deleted 🔖 Link in the comments 👇

draven

38,752 views • 23 days ago

THIS IS F*CKING BRILLIANT. Anthropic’s docs describe one of the smartest ways to run Opus three models work on one task and each does only the part it’s cheapest at. start Claude Code like this: claude -> model sonnet -> advisor opus you stop paying Opus prices for work that doesn’t need Opus > Sonnet runs the session and edits the code > Haiku 5.5 searches the repo and docs, at up to 40x less per token than Opus 5.5 > Opus waits as the advisor and reads the whole session, tool calls included Claude usually calls Opus at three points: > before it locks a plan > when the same error comes back > right before it says it’s done Sonnet drives. Haiku scouts. Opus signs off and if you go one step lower, most of an agent’s day is tiny choices like picking a file or a tool, or deciding whether to retry. Jev takes those first, so Sonnet and Opus only see the moments that are really open the whole setup Sonnet runs the main session Explore on Haiku 5.5 searches and pulls docs, read-only Opus on call as the advisor give Claude Code this setup and the prompt below ↓ “Set up my Claude Code exactly like this: Main model Sonnet, advisor Opus, saved as advisorModel in my settings. One read-only subagent named exactly Explore, on claude-haiku-5-5, to replace the built-in one. One line in CLAUDE.md: ask the advisor before locking a plan, after the same error twice, and before calling a task done, not more often. Check what could turn the advisor off (CLAUDE_CODE_DISABLE_ADVISOR_TOOL, DISABLE_TELEMETRY) and only report it. Show every change as a diff first. Don’t edit anything until I say go.” before you turn it on: every advisor call re-reads the whole transcript with no cache. on a 200k session that’s about $0.80 in input, against $0.04 for a normal request. keep it to a few calls the built-in Explore agent runs on your main model unless a subagent is named exactly Explore. any other name does nothing turning the advisor on or off keeps your prompt cache. switching models rebuilds it it needs the Anthropic API, and it stays off if telemetry is disabled bookmark it before your next long session
0:17

THIS IS F*CKING BRILLIANT. Anthropic’s docs describe one of the smartest ways to run Opus three models work on one task and each does only the part it’s cheapest at. start Claude Code like this: claude -> model sonnet -> advisor opus you stop paying Opus prices for work that doesn’t need Opus > Sonnet runs the session and edits the code > Haiku 5.5 searches the repo and docs, at up to 40x less per token than Opus 5.5 > Opus waits as the advisor and reads the whole session, tool calls included Claude usually calls Opus at three points: > before it locks a plan > when the same error comes back > right before it says it’s done Sonnet drives. Haiku scouts. Opus signs off and if you go one step lower, most of an agent’s day is tiny choices like picking a file or a tool, or deciding whether to retry. Jev takes those first, so Sonnet and Opus only see the moments that are really open the whole setup Sonnet runs the main session Explore on Haiku 5.5 searches and pulls docs, read-only Opus on call as the advisor give Claude Code this setup and the prompt below ↓ “Set up my Claude Code exactly like this: Main model Sonnet, advisor Opus, saved as advisorModel in my settings. One read-only subagent named exactly Explore, on claude-haiku-5-5, to replace the built-in one. One line in CLAUDE.md: ask the advisor before locking a plan, after the same error twice, and before calling a task done, not more often. Check what could turn the advisor off (CLAUDE_CODE_DISABLE_ADVISOR_TOOL, DISABLE_TELEMETRY) and only report it. Show every change as a diff first. Don’t edit anything until I say go.” before you turn it on: every advisor call re-reads the whole transcript with no cache. on a 200k session that’s about $0.80 in input, against $0.04 for a normal request. keep it to a few calls the built-in Explore agent runs on your main model unless a subagent is named exactly Explore. any other name does nothing turning the advisor on or off keeps your prompt cache. switching models rebuilds it it needs the Anthropic API, and it stays off if telemetry is disabled bookmark it before your next long session

rvaniaaa

62,223 views • 1 month ago