Happy New Year! In December, I shared 31 days of Claude Code...

                            1
                        
Happy New Year!

In December, I shared 31 days of Claude Code features. Thanks for following along and the replies, the questions, the "wait it can do that?"

If you missed any (or want them all in one place), here it is.

Your 2026 terminal is about to be very different. 🧵

2

! to execute bash commands instantly and inject the output into context:

Ado@adocomplete

Advent of Claude Day 1 - The ! Prefix

Don't waste tokens asking "can you run git status or tests?"

Just type: ! followed by your bash command

The ! prefix executes bash instantly and injects the output into context.

No model processing. No delay. No wasted tokens.

3

Double tap Esc to rewind:

Ado@adocomplete

Advent of Claude Day 2 - Double Esc to Rewind

Want to try a “what if we…” approach?

Go wild → if it gets weird → Esc Esc to jump back to a clean checkpoint and revise.

You can rewind the conversion, code, or both.

4

How to work with extended thinking in the Claude API:

Ado@adocomplete

Advent of Claude Day 3 - Extended Thinking

Want Claude to actually think before answering when using the API?

Just add:

thinking: { type: "enabled", budget_tokens: 5000 }

Claude shows its step-by-step reasoning in thinking blocks before responding.

5

Session management in Claude Code:

Ado@adocomplete

Advent of Claude Day 4 - Session Management

Accidentally closed your terminal? Laptop died? No problem.

claude --continue → picks up your last conversation instantly
claude --resume → shows a picker to choose any past session

Context preserved. Momentum restored.

6

# to make updates to ClaudeMD files.

Note: This was removed recently, you can just ask Claude now to remember details or update ClaudeMD file directly:

Ado@adocomplete

Advent of Claude Day 5 - The # Prefix

Want to save something to Claude's memory without opening a file?

Just start your message with #, for example:

# Always use bun instead of npm

Claude asks where to save it. Memory updated. Keep coding.

7

Vim mode in Claude Code:

Ado@adocomplete

Advent of Claude Day 6 - Vim mode

Tired of reaching for the mouse to edit your prompts?

Just type: /vim

Now you get full vim-style editing in Claude Code:

h j k l → navigate
ciw → change word
dd → delete line
w b → jump by word

Edit prompts at the speed of thought.

8

Run Claude Code in headless mode:

Ado@adocomplete

Advent of Claude Day 7 - Headless Mode

Want to use Claude Code in scripts, CI/CD, or automation?

Just add -p:

* claude -p "Fix the lint errors"
* claude -p "List all the functions " | grep "async"
* git diff | claude -p "Explain these changes"

AI in your pipeline.

9

/context to better understand what is consuming your tokens:

Ado@adocomplete

Advent of Claude Day 8 - The /context Command

Want to see what's eating your context window?

Type /context to see exactly what's consuming your token window including:

* The system prompt
* MCP servers
* Memory files
* and more

It's like having X-ray vision for your tokens.

10

Claude Code remote:

Ado@adocomplete

Advent of Claude Day 9 - Claude Code Remote

& your prompt
↳ Sends the task to Claude Code on the web (runs in background)

claude --teleport session_abc123
↳ Pulls and resumes the session in your local terminal to continue working

Claude at home and on the go.

11

/stats to see how you're using Claude Code:

Ado@adocomplete

Advent of Claude Day 10 - /stats

2023: "Check out my GitHub contribution graph"
2025: "Check out my Claude Code stats"

Just type /stats

Orange is the new green.

12

Named sessions for easier management (pair it with session management to quickly get back into the groove):

Ado@adocomplete

Advent of Claude Day 11 - Named Sessions

Claude Code now lets you name sessions.

/rename api-migration
→ Names your current session

/resume api-migration or claude --resume api-migration
→ Resumes the session

Your branches have names. Sessions should too.

13

Ultrathink to get more done:

Note: Typing Ultrathink still works, but the other thinking modes have been removed. See the thread for more info on how Claude Code thinking modes work now.

Ado@adocomplete

Advent of Claude Day 12 - Ultrathink

You can control how hard claude will think before giving you a response.

"think" → 4k thinking tokens
"think hard" → 10k thinking tokens
"ultrathink" → 31,999 thinking tokens

Just say the magic word anywhere in your prompt.

14

Stash your prompts:

Ado@adocomplete

Advent of Claude Day 13 - Prompt Stashing

It's like git stash, but for your prompt.

Ctrl+S saves your draft. Send something else. Draft auto-restores.

No more copying to a scratchpad. No more losing your train of thought.

15

Customize your Claude Code status line with info you care about:

Ado@adocomplete

Advent of Claude Day 14 - /statusline

Claude Code has a customizable status bar.

/statusline puts info relevant to you at the bottom of Claude Code. Customize it however you want!

Git info, model, tokens, weather, Bitcoin price, prayer times, Wordle streak - it's your call.

16

Claude on the edge with --dangerously-skip-permissions:

Ado@adocomplete

Advent of Claude Day 15 - YOLO Mode

Tired of Claude Code asking permission for everything?

claude --dangerously-skip-permissions

This flag says yes to everything. It has "danger" in the name, so please use it wisely.

17

Bring determinism to Claude Code with hooks:

Ado@adocomplete

Advent of Claude Day 16 - Hooks

Hooks are shell commands that run at predetermined lifecycle events.

Run /hooks or configure in .claude/settings.json.

Use hooks to block dangerous commands, provide notifications, and much more.

Deterministic control over probabilistic AI.

18

Set the right context with the right files with @ mentions:

Ado@adocomplete

Advent of Claude Day 17 - @ Mentions

@ mentions in Claude Code may be overpowered:

* Add one or many files into context instantly
* Quickly list directories
* Enable/disable/invoke MCP servers and subagents

@ is the fastest path from "I need context" to "Claude has context."

19

Claude code can now suggest follow-up prompts:

Ado@adocomplete

Advent of Claude Day 18 - Prompt Suggestions

Claude can predict what you'll ask next

Finish a task → see a grayed-out follow-up suggestion appear:

Tab → accept and edit
Enter → accept and run immediately

Tab used to autocomplete your code. Now it autocompletes your work.

20

Claude Code🤝 Chrome:

Ado@adocomplete

Advent of Claude Day 19 - Claude Code + Chrome

Claude can now directly interact with your browser.

Navigating pages, clicking buttons, reading console errors, inspecting the DOM, and more.

"Fix the bug and verify it works" is now one prompt.

/chrome to start.

21

Teach Claude Code how to do specialized tasks with Skills:

Ado@adocomplete

Advent of Claude Day 20 - Agent Skills

Skills = folders of instructions, scripts, and resources that teach Claude how to do specialized tasks.

Packaged once. Usable everywhere. And since Agent skills are now an open standard, I mean everywhere!

22

Easily share your Claude Code workflows with Plugins:

Ado@adocomplete

Advent of Claude Day 21 - Plugins

"Can you send me your Claude Code setup?"

sends 47 files across 12 directories

"...nevermind"

That era is over

/plugin install my-setup

Discover new workflows via marketplace

Commands, agents, skills, hooks, MCP. One package. One command.

23

Reverse search to find past prompts:

Ado@adocomplete

Advent of Claude Day 22 - Ctrl + R (Reverse Search)

Your past prompts are searchable within Claude Code.

Ctrl+R → reverse search through your prompt history
Ctrl+R again → cycle through matches
Enter → run it
Tab → edit first

Don't retype. Recall.

24

Onboard Claude Code to your codebase quickly with /init:

Ado@adocomplete

Advent of Claude Day 23 - /init

Everybody needs onboarding docs. With /init, Claude writes its own.

/init → Claude reads your codebase and generates a CLAUDEmd file with build commands, test instructions, key directories, conventions.

Claude onboards itself. Good Claude.

25

/export your Claude Code threads:

Ado@adocomplete

Advent of Claude Day 24 - /export

Sometimes you want receipts.

/export → dumps your entire conversation to markdown including every prompt, every response, every tool call.

For documentation. For training. For proving to your past self that yes, you did try that already.

26

Split the work with Subagents:

Ado@adocomplete

Advent of Claude Day 25 - Subagents

Santa doesn't wrap every gift. He has elves.

Subagents are Claude's elves.

Each subagent gets it's own 200k context window, performs specialized tasks, runs in parallel and merges it's output back to main agent.

Delegate like Santa. 🎅

27

Save any prompt as a reusable command:

Ado@adocomplete

Advent of Claude Day 26 - Commands

Save any prompt as a reusable command.

Single markdown file in. Slash command out. Accepts arguments too.

/explain $ARGUMENTS → /explain src/auth.ts

Stop repeating yourself.

28

Define boundaries with Sandbox mode:

Ado@adocomplete

Advent of Claude Day 27 - Sandbox Mode

"Can I run npm install?" [Allow]
"Can I run npm test?" [Allow]
"Can I cat this file?" [Allow]
×100

/sandbox → boundaries set → Claude gets to work.

Define boundaries once. Claude works freely inside them.

YOLO speed. Actual security.

29

Check your /usage and get more with /extra-usage:

Ado@adocomplete

Advent of Claude Day 28 - /usage

"Am I about to hit my limit?"

/usage → check your limits
/extra-usage → remove your limits

Know your limits. Then exceed them.

30

Plan before you execute:

Ado@adocomplete

Advent of Claude Day 29 - Plan mode

Clear the fog of war first.

Shift+Tab twice → Plan mode.

Claude can read, search, analyze and explore your codebase but won't edit anything.

Think twice. Execute once.

Seriously though, I default to plan mode 90% of the time.

31

Language Server Protocol support in your Claude Code terminal:

Ado@adocomplete

Advent of Claude Day 30 - LSP

Language Server Protocol in Claude Code provides

- Instant diagnostics to see errors and warnings
- Navigation with go to definition, find references, and more
- Type info and docs for code symbols

Claude Code now has IDE-level code intelligence

32

And finally, the Claude Agent SDK to build your own Claude Code and more:

Ado@adocomplete

Advent of Claude Day 31 - Claude Agent SDK

I saved the best for last.

The same agent loop, tools, and context management that power Claude Code is available as an SDK.

Build agents that work like Claude Code in as little as 10 lines of code.

This is just the beginning. 🎇🥂

                            33
                        
That's all 31.

If 2025 was the year that AI really learned to code well, I predict 2026 will be the year we learn to better build with it.

Thank you again for tuning in and happy building!

Happy New Year! In December, I shared 31 days of Claude Code...

Actions

What You Can Do

Actions

What You Can Do

More in AI & Machine Learning

More from @adocomplete