karthiks

#!/bin/bash

# This is my ClaudeCode's StatusLine bashscript for showing the current directory, model name, and git status if applicable.
# ref.: https://code.claude.com/docs/en/statusline

# Read JSON data that Claude Code sends to stdin
input=$(cat)

# Extract information from JSON using jq
CC_VER=$(echo "$input" | jq -r '.version')

karthiks

# 1. Handle line endings automatically for text files
# This ensures LF in the repo, but OS-native endings in the working directory
* text=auto

# 2. Explicitly declare text files (especially your programming language file endings) to be normalized to LF in the repo
*.c       text
*.h       text
*.cpp     text
*.py      text
*.java     text

karthiks

/help #shows complete command list
/help compact #shows detailed syntax for the compact command

/init #scans package.json, tsconfig.json, etc., and creates a 15-30 line CLAUDE.md file with detected conventions
/doctor #runs 5-8 checks in under 15 seconds and reports any issues with green (OK) or red (failure) status

/compact #compresses current session
/clear #use when switching topics or after 20+ exchanges; keyboard shortcut Ctrl+L
/rewind #runs once to undo one turn; can be run multiple times to go back further

karthiks

The Two Pillars of Memory

Claude Code uses a dual-memory system to balance your explicit instructions with its own learned observations.

Feature	CLAUDE.md (Static Memory)	MEMORY.md (Auto Memory)
Who Writes It?	You (Manual)	Claude (Automated)
Content	"Project rules, architecture, tech stack."	"Patterns, ""gotchas,"" and user preferences."
Location	Your Project Root (git-tracked).	~/.claude/projects/ (system-local).
Persistence	Permanent and shared with your team.	Local to your machine; session-dependent.

karthiks

Session Context Window (200k total)
┌────────────────────────────────────────────┐
│ System Prompt + Tools: 15.9k (8%)          │ ← Fixed allocation
│ CLAUDE.md + Memory: 2.8k (1.4%)            │ ← Fixed allocation
├────────────────────────────────────────────┤
│ Messages (grows over time): 89.5k (44.7%)  │ ← This grows
├────────────────────────────────────────────┤
│ Free Space: 59k (29.5%)                    │ ← This shrinks
│ Autocompact Buffer: 33k (16.5%) │ ← Reserved

karthiks

Here's a practical reference table for what healthy, warning, and critical levels look like for each category:

Category	Healthy	Warning	Critical	Breach Consequences
System Prompt	<5k tokens	5-8k tokens	>8k tokens	Reduced context for actual work; slower responses
System Tools	<10k tokens	10-20k tokens	>20k tokens	Context starvation; frequent compacting; tool failures
Memory Files	<1k tokens	1-2k tokens	>2k tokens	Important instructions being truncated; Claude "forgetting" conventions
Messages	<40% of total	40-60% of total	>60% of total	Rapidly approaching limit; autocompact ineffective
Free Space	>30%	15-30%	<15%	Autocompact may fail; Claude may refuse new tasks
Autocompact Buffer	>20%	15-20%	<15%	Summaries become poor; important context lost
Total Usage	<70%	70-85%	>85%	Performance degradation; risk of hitting limit mid-task

karthiks

flowchart TB
    subgraph Request1[Request 1: Cache Write]
        A1[System prompt<br/>CLAUDE.md + Tools] 
        A2[User message 1]
        A3[Cache Storage]
        
        A1 --> A3
        A2 --> B[Process full request]
 A3 --&gt; B

karthiks

### Check out detailed blog post at https://blog.codonomics.com/2026/01/accessing-windows-host-from-wsl2-linux.html 
### Start: Adding Windows Host IP to WSL2 Hosts File
# Add to the end of ~/.bashrc or ~/.zshrc
echo "Updating Windows host IP in /etc/hosts..."

# Get Windows IP from default route
WIN_HOST_IP=$(ip route show | grep -i default | awk '{print $3}')

# Check if we got a valid IP
if [ -n "$WIN_HOST_IP" ]; then

karthiks

# =============================================
# MINICONDA CHEATSHEET (Python Environment & Package Management)
# =============================================
# Lightweight version of Anaconda for Python environments.
# Install: https://docs.conda.io/projects/miniconda/en/latest/
# Docs: https://conda.io/projects/conda/en/latest/commands.html

# ---------------------------------------------------------------------
# 1. INSTALLATION & SETUP
# ---------------------------------------------------------------------

karthiks

# =============================================
# UV CHEATSHEET (Python Package & Env Management)
# =============================================
# Astral's ultra-fast alternative to pip/venv.
# Install: curl -LsSf https://astral.sh/uv/install.sh | sh
# Docs: https://github.com/astral-sh/uv

# ---------------------------------------------------------------------
# 1. VIRTUAL ENVIRONMENTS
# ---------------------------------------------------------------------