Usage Examples

# Simple questions
ask "what's the difference between TCP and UDP?"

# Code explanations
ask "explain what this regex does: ^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$"

# Best practices
ask "what's the best way to handle errors in bash scripts?"

# Analyze logs
tail -100 /var/log/nginx/access.log | ask "summarize the traffic patterns"

# Debug errors
cat error.log | ask "what's causing these errors and how to fix?"

# Process data
ps aux | ask "what processes are using the most memory?"

# Security expert persona
ask --system "You are a security expert" "review this code for vulnerabilities" < auth.py

# Database performance expert
ask --system "You are a database performance expert" "optimize this SQL query"

# Technical writer
ask --system "You are a technical documentation writer" "document this API" < routes.js

# Creative writing (higher temperature = more creative)
ask -t 1.5 "write a creative story about a developer"

# Deterministic output (lower temperature = more focused)
ask -t 0.1 "extract structured data from this JSON" < data.json

# Balanced (default is 1.0)
ask -t 1.0 "explain how async/await works"

# For programmatic use
ask --json --no-stream "list top 3 security practices" | jq -r '.content[0].text'

# Batch processing
for file in *.log; do
  errors=$(ask --json --no-stream "count error lines" < "$file" | jq -r '.count')
  echo "$file: $errors errors"
done

# Complex queries across multiple lines
ask --multiline <<'EOF'
Given this error stack trace, explain:
1. Root cause
2. How to fix
3. How to prevent

[paste stack trace here]
EOF

# Find files intelligently
ask "find all javascript files modified in the last week"

# Analyze disk usage
du -sh * | ask "which directories should I clean up first?"

# Permission issues
ask "why can't I write to this directory?" --context full

# Debug hanging processes
ps aux | grep myapp | ask "is this process stuck?"

# Resource monitoring
top -bn1 | ask "diagnose performance issues"

# Kill processes safely
ask --agent "kill all zombie processes"

# Test connectivity
ping -c 5 google.com | ask "analyze the connection quality"

# Port debugging
netstat -tulpn | ask "which service is using port 8080?"

# SSL certificates
openssl s_client -connect example.com:443 | ask "when does this cert expire?"

# Smart commits (requires staged changes)
git add .
ask commit
# Generates: "feat(api): add user authentication endpoint"

# Review changes
git diff main..feature | ask "review this PR for issues"

# PR review command
git checkout feature-branch
ask pr-review
# Reviews current branch against main

# Find commits
git log --oneline | ask "find the commit that added authentication"

# Review specific file
git diff main src/auth.js | ask "security review this authentication code"

# Find bugs
cat bug_report.txt | ask "what's the root cause?" --context full

# With security expert
ask --system "You are a security auditor" "find vulnerabilities" < app.py

# Generate tests
ask --fn test_user_login "create pytest tests for user login function"

# Debug test failures
pytest -v | ask "why are these tests failing?"

# Coverage analysis
coverage report | ask "what critical paths are missing tests?"

# Diagnose last command failure
make build
# [build fails]
ask diagnose

# With full context
npm test
# [tests fail]
ask --context full diagnose

# Simple task
ask --agent "create a backup directory with today's date"

# Auto-approve low/medium risk (recommended)
ask --agent "organize files by type"
# Type 'a' when prompted to auto-approve safe commands

# Dry run (show plan without executing)
ask --agent --dry-run "delete old log files"

ask --agent "organize these files by file type into subdirectories"
ask --agent "find and remove files larger than 100MB not modified in 30 days"
ask --agent "create timestamped backup of all .conf files"
ask --agent --dry-run "rename all JPEG files to lowercase"

ask --agent "remove all stopped containers and unused images"
ask --agent "restart unhealthy containers"
ask --agent "find containers using more than 1GB RAM"
ask --agent --tools "docker" "cleanup unused volumes"

# Only allow specific tools
ask --agent --tools "curl,jq" "fetch and parse this API endpoint"

# Git-only operations
ask --agent --tools "git" "create feature branch and initial commit"

# Safe read-only tools
ask --agent --tools "cat,grep,find" "analyze log patterns"

# Generate a function
ask --fn parse_nginx "extract 500 errors from nginx access logs"

# ask will:
# 1. Generate the function
# 2. Validate syntax
# 3. Show preview
# 4. Ask: Save? (y/n/e to edit)

ask --fn parse_nginx_errors "extract all 500 errors with timestamp and URL"
ask --fn find_slow_queries "find SQL queries taking more than 1 second"
ask --fn tail_errors "tail system logs and highlight errors in red"
ask --fn parse_json_logs "parse and pretty-print JSON log entries"

# Source the functions file
source ~/.config/ask/functions.sh

# Now use your functions
parse_nginx_errors /var/log/nginx/access.log
csv_unique data.csv email
git_clean_branches

# Add to .bashrc for permanent use
echo 'source ~/.config/ask/functions.sh' >> ~/.bashrc

# Start interactive session
ask

# You'll see the banner and prompt
ask> your question here

ask> /clear     # Clear conversation history
ask> /save      # Save conversation
ask> /load      # Load previous conversation
ask> /models    # List available models
ask> /switch openai gpt-4o  # Switch provider/model
ask> /context full          # Change context level
ask> /help      # Show help
ask> /exit      # Exit

cat sales.csv | ask "what are the top 5 products by revenue?"
cat metrics.csv | ask "detect any unusual patterns in this data"
cat data.csv | ask "create a summary report with key insights"

cat api_response.json | ask "extract all user emails where status is active"
ask "is this valid JSON and what structure does it have?" < data.json
cat input.json | ask "convert this to CSV format"

cat app.log | ask "group errors by type and show frequency"
cat access.log | ask "calculate average response time by endpoint"
cat auth.log | ask "find suspicious login attempts"

ask --agent "find files with world-writable permissions"
ask "check if these passwords meet security requirements" < passwords.txt
npm audit | ask "prioritize these vulnerabilities by severity"
ask --system "You are a security expert" "audit this code" < app.py

ask "scan this code for GDPR compliance issues" < user_service.py
ask --agent "list all open source licenses in node_modules"
ask --agent "scan for exposed API keys or credentials"

kubectl get pods | ask "which pods are crashlooping?"
kubectl top nodes | ask "which node is overutilized?"
kubectl logs -f mypod | ask "alert me to errors in real-time"
kubectl get deployment myapp -o yaml | ask "check for security issues"

ask "why did this build fail?" < build.log
ask --context full "debug this deployment failure"
ask "compare these benchmark results" < benchmarks.txt
ask --json --no-stream "extract failure reason" < ci.log | jq -r '.reason'

terraform plan | ask "explain these changes in plain English"
ask "convert this bash script to an Ansible playbook" < deploy.sh
ask "review this nginx config for security issues" < nginx.conf

# Expert persona + full context + agent mode
ask --system "You are a DevOps expert" \
    --context full \
    --agent \
    "set up monitoring for this service"

# JSON output + no streaming for scripting
ask --json --no-stream \
    --system "You are a data analyst" \
    "summarize key metrics" < metrics.csv | jq

# Set default provider
export ASK_PROVIDER=openai
ask "your question"  # Uses OpenAI

# Set default model
export ASK_MODEL=gpt-4o

# Temporary override
ASK_MODEL=gpt-4o-mini ask "quick question"

ask --context min "quick question"     # Fast, minimal context
ask --context auto "debug this"        # Balanced (default)
ask --context full "complex analysis"  # Maximum context

git diff | ask "summarize changes" | ask "suggest commit message"
find . -name "*.js" | xargs cat | ask "find code smells"

ask -m gpt-4o-mini "quick question"        # Fast & cheap
ask -m claude-opus-4 "design architecture"  # Powerful
ask --list-models                            # List all

ask "explain: find . -type f -name '*.log' -mtime +30 -delete"
ask "how does parameter expansion work in bash?"
ask "explain grep options with examples"

cat myfunction.py | ask "generate docstring for this function"
ask "create a README for this project" --context full
ask "document this REST API" < api_routes.js

ask --context full "daily system health check"
git status | ask "summarize what I was working on"
ask "based on my recent commits, what should I focus on today?"

git diff main..feature > /tmp/pr.diff
ask --system "You are a senior code reviewer" "review this PR" < /tmp/pr.diff
ask --system "You are a security expert" "security review" < /tmp/pr.diff
ask "suggest test cases for this PR" < /tmp/pr.diff

./myapp 2>&1 | tee error.log
ask "what's wrong?" < error.log
ask --context full "debug this error"
ask --agent --dry-run "fix this issue"
ask --agent "fix this issue"

ask keys set anthropic
ask keys set openai
ask keys set openrouter
ask keys list

ask keys list
ask keys remove openai
ask keys path
ANTHROPIC_API_KEY='different-key' ask "test query"

which ask
echo $PATH
export PATH="$HOME/.local/bin:$PATH"

ask keys list
ask keys set anthropic
ask "test"

ask -n "your query"
export ASK_STREAM=false

ask --context min "quick question"
ask -m gpt-4o-mini "simple task"
ask -n "query"