Jq

jq — JSON Querying and Transformation

Overview

jq is the standard CLI tool for querying and reshaping JSON. This skill covers practical, expert-level usage: filtering deeply nested data, transforming structures, aggregating values, and composing jq into shell pipelines. Every example is copy-paste ready for real workflows.

When to Use This Skill

• Use when parsing JSON output from APIs, CLI tools (AWS, GitHub, kubectl, docker), or log files
• Use when transforming JSON structure (rename keys, flatten arrays, group records)
• Use when the user needs jq inside a bash script or one-liner
• Use when explaining what a complex jq expression does

How It Works

jq takes a filter expression and applies it to JSON input. Filters compose with pipes (|), and jq handles arrays, objects, strings, numbers, booleans, and null natively.

Basic Selection

# Extract a field
echo '{"name":"alice","age":30}' | jq '.name'
# "alice"

# Nested access
echo '{"user":{"email":"a@b.com"}}' | jq '.user.email'

# Array index
echo '[10, 20, 30]' | jq '.[1]'
# 20

# Array slice
echo '[1,2,3,4,5]' | jq '.[2:4]'
# [3, 4]

# All array elements
echo '[{"id":1},{"id":2}]' | jq '.[]'

Filtering with select

# Keep only matching elements
echo '[{"role":"admin"},{"role":"user"},{"role":"admin"}]' \
  | jq '[.[] | select(.role == "admin")]'

# Numeric comparison
curl -s https://api.github.com/repos/owner/repo/issues \
  | jq '[.[] | select(.comments > 5)]'

# Test a field exists and is non-null
jq '[.[] | select(.email != null)]'

# Combine conditions
jq '[.[] | select(.active == true and .score >= 80)]'

Mapping and Transformation

# Extract a field from every array element
echo '[{"name":"alice","age":30},{"name":"bob","age":25}]' \
  | jq '[.[] | .name]'
# ["alice", "bob"]

# Shorthand: map()
jq 'map(.name)'

# Build a new object per element
jq '[.[] | {user: .name, years: .age}]'

# Add a computed field
jq '[.[] | . + {senior: (.age > 28)}]'

# Rename keys
jq '[.[] | {username: .name, email_address: .email}]'

Aggregation and Reduce

# Sum all values
echo '[1, 2, 3, 4, 5]' | jq 'add'
# 15

# Sum a field across objects
jq '[.[].price] | add'

# Count elements
jq 'length'

# Max / min
jq 'max_by(.score)'
jq 'min_by(.created_at)'

# reduce: custom accumulator
echo '[1,2,3,4,5]' | jq 'reduce .[] as $x (0; . + $x)'
# 15

# Group by field
jq 'group_by(.department)'

# Count per group
jq 'group_by(.status) | map({status: .[0].status, count: length})'

String Interpolation and Formatting

# String interpolation
jq -r '.[] | "\(.name) is \(.age) years old"'

# Format as CSV (no header)
jq -r '.[] | [.name, .age, .email] | @csv'

# Format as TSV
jq -r '.[] | [.name, .score] | @tsv'

# URL-encode a value
jq -r '.query | @uri'

# Base64 encode
jq -r '.data | @base64'

Working with Keys and Paths

# List all top-level keys
jq 'keys'

# Check if key exists
jq 'has("email")'

# Delete a key
jq 'del(.password)'

# Delete nested keys from every element
jq '[.[] | del(.internal_id, .raw_payload)]'

# Recursive descent: find all values for a key anywhere in tree
jq '.. | .id? // empty'

# Get all leaf paths
jq '[paths(scalars)]'

Conditionals and Error Handling

# if-then-else
jq 'if .score >= 90 then "A" elif .score >= 80 then "B" else "C" end'

# Alternative operator: use fallback if null or false
jq '.nickname // .name'

# try-catch: skip errors instead of halting
jq '[.[] | try .nested.value catch null]'

# Suppress null output with // empty
jq '.[] | .optional_field // empty'

Practical Shell Integration

# Read from file
jq '.users' data.json

# Compact output (no whitespace) for further piping
jq -c '.[]' records.json | while IFS= read -r record; do
  echo "Processing: $record"
done

# Pass a shell variable into jq
STATUS="active"
jq --arg s "$STATUS" '[.[] | select(.status == $s)]'

# Pass a number
jq --argjson threshold 42 '[.[] | select(.value > $threshold)]'

# Slurp multiple JSON lines into an array
jq -s '.' records.ndjson

# Multiple files: slurp all into one array
jq -s 'add' file1.json file2.json

# Null-safe pipeline from a command
kubectl get pods -o json | jq '.items[] | {name: .metadata.name, status: .status.phase}'

# GitHub CLI: extract PR numbers
gh pr list --json number,title | jq -r '.[] | "\(.number)\t\(.title)"'

# AWS CLI: list running instance IDs
aws ec2 describe-instances \
  | jq -r '.Reservations[].Instances[] | select(.State.Name=="running") | .InstanceId'

# Docker: show container names and images
docker inspect $(docker ps -q) | jq -r '.[] | "\(.Name)\t\(.Config.Image)"'

Advanced Patterns

# Transpose an object of arrays to an array of objects
# Input: {"names":["a","b"],"scores":[10,20]}
jq '[.names, .scores] | transpose | map({name: .[0], score: .[1]})'

# Flatten one level
jq 'flatten(1)'

# Unique by field
jq 'unique_by(.email)'

# Sort, deduplicate and re-index
jq '[.[] | .name] | unique | sort'

# Walk: apply transformation to every node recursively
jq 'walk(if type == "string" then ascii_downcase else . end)'

# env: read environment variables inside jq
export API_KEY=secret
jq -n 'env.API_KEY'

Best Practices

• Always use -r (raw output) when passing jq results to shell variables or other commands to strip JSON string quotes
• Use --arg / --argjson to inject shell variables safely — never interpolate shell variables directly into filter strings
• Prefer map(f) over [.[] | f] for readability
• Use -c (compact) for newline-delimited JSON pipelines; omit it for human-readable debugging
• Test filters interactively with jq -n and literal input before embedding in scripts
• Use empty to drop unwanted elements rather than filtering to null

Security & Safety Notes

• jq is read-only by design — it cannot write files or execute commands
• Avoid embedding untrusted JSON field values directly into shell commands; always quote or use --arg

Common Pitfalls

• Problem: jq outputs null instead of the expected value Solution: Check for typos in key names; use keys to inspect actual field names. Remember JSON is case-sensitive.

• Problem: Numbers are quoted as strings in the output Solution: Use --argjson instead of --arg when injecting numeric values.

• Problem: Filter works in the terminal but fails in a script Solution: Ensure the filter string uses single quotes in the shell to prevent variable expansion. Example: jq '.field' not jq ".field".

• Problem: add returns null on an empty array Solution: Use add // 0 or add // "" to provide a fallback default.

• Problem: Streaming large files is slow Solution: Use jq --stream or switch to jstream/gron for very large files.

Related Skills

• @bash-pro — Wrapping jq calls in robust shell scripts
• @bash-linux — General shell pipeline patterns
• @github-automation — Using jq with GitHub CLI JSON output

🧠 AGI Framework Integration

Adapted for @techwavedev/agi-agent-kit Original source: antigravity-awesome-skills

Qdrant Memory Integration

Before executing complex tasks with this skill:

python3 execution/memory_manager.py auto --query "<task summary>"

• Cache hit? Use cached response directly — no need to re-process.
• Memory match? Inject context_chunks into your reasoning.
• No match? Proceed normally, then store results:

python3 execution/memory_manager.py store \\
  --content "Description of what was decided/solved" \\
  --type decision \\
  --tags jq <relevant-tags>

Agent Team Collaboration

• This skill can be invoked by the orchestrator agent via intelligent routing.
• In Agent Teams mode, results are shared via Qdrant shared memory for cross-agent context.
• In Subagent mode, this skill runs in isolation with its own memory namespace.

Local LLM Support

When available, use local Ollama models for embedding and lightweight inference:

• Embeddings: nomic-embed-text via Qdrant memory system
• Lightweight analysis: Local models reduce API costs for repetitive patterns