1️⃣ Step 1: Run Multiple Sessions at the Same Time
Most people run one Claude session and wait for it to finish before starting the next task. Boris runs 10 to 15 at once.
How to set it up
2️⃣ Step 2: Install Your CLAUDE.md File
Go to claude.ai/code and open additional sessions in Chrome running alongside your terminal sessions.
→ Hand off a terminal session to the browser using & in your terminal.
→ Start sessions from your phone in the morning and pick them up on desktop later.
3️⃣ Step 3: Use Opus 4.5 With Thinking for Everything
Boris uses Opus 4.5 with thinking for every task. Even though it is slower per response, you steer it less and it is better at using tools, so it is almost always faster overall.
4️⃣ Step 4: Set Up Your CLAUDE.md and Check It Into Git
This is the most important file. It lives in your project root and Claude reads it at the start of every session.
→ The whole team shares one CLAUDE.md checked into git.
→ Every time Claude does something wrong, the rule goes in the file so it never happens again.
After every correction, end your message with:
5️⃣ Step 5: Tag Claude in Code Reviews to Update CLAUDE.md
When reviewing a teammate's code, tag @.claude on the pull request to add new rules to CLAUDE.md as part of the review.
→ Install the Claude Code GitHub action using /install-github-action to enable this.
6️⃣ Step 6: Always Start in Plan Mode
A good plan is the most important thing.
→ Press Shift and Tab twice at the start of any session to enter Plan Mode.
→ Go back and forth with Claude until the plan is exactly right.
→ Switch to auto-accept edits mode and Claude implements it in one go without back and forth.
7️⃣ Step 7: Set Up Slash Commands for Every Repeated Task
If you do something more than once a day, make it a slash command.
→ Boris uses /commit-push-pr dozens of times every day. It handles saving, submitting, and opening a review for your code in one command.
→ Commands live in .claude/commands/ and are checked into git so everyone on your team uses them.
8️⃣ Step 8: Install Your Subagents
Boris uses three subagents on almost every piece of work.
Three core subagents
Install them from the ChernyCode repo included in this resource.
9️⃣ Step 9: Add a PostToolUse Hook to Format Code
Add a PostToolUse hook that runs your formatter every time Claude edits a file.
Claude generates well-formatted code most of the time. The hook handles the remaining cases before they cause errors later.
🔟 Step 10: Set Safe Permissions Instead of Skipping Them
Never use --dangerously-skip-permissions.
→ Run /permissions and pre-allow the specific commands you know are safe.
→ Check .claude/settings.json into git so the whole team shares the same safe defaults.
Step 11: Connect Claude to Your Tools via MCP
Claude Code can search and post to Slack, run BigQuery queries for analytics, and pull error logs from Sentry.
→ Check your .mcp.json into git so the whole team gets the same connections.
Step 12: Handle Long Tasks With Background Verification
For tasks that run a long time, either prompt Claude to verify its work with a background agent when done, use an agent Stop hook to do it automatically, or use the ralph-wiggum plugin.
Step 13: Always Give Claude a Way to Verify Its Own Work
Boris says this is the single most important thing.
→ Claude tests every change using the Chrome extension. It opens a browser, tests the UI, and keeps iterating until the code works and the experience feels right.
→ Add this to the end of any session: "Open a browser, test this end to end, and fix anything that does not work."
→ A verification loop produces 2 to 3 times better output than no verification.
Using Prosp.ai with Claude to book 15 calls/week from LinkedIn
1:1 with a GTM engineer
