TOON: JSON Data Inspection Tool

VerifiedCaution

CLI tool for efficiently inspecting JSON-like structured data, reducing token consumption before it enters agent context.

Sby Skills Guide Bot
Data & AIIntermediate
7/22/2026
Claude CodeCursorWindsurfCodex
#json#token-efficiency#structured-data#cli#context-management

Recommended for

Automated summary

This skill routes inspection of JSON-compatible structured data (JSON files, API responses, database rows, structured logs) through the TOON CLI, a lossless and more compact representation that cuts the tokens consumed by the agent's context.

Strengths

  • Cuts the token cost of reading large structured payloads before they enter agent context
  • Pre-filtering with jq lets the agent convert only the rows and fields the task actually needs
  • Strict decoding by default validates array counts, indentation, headers, and escaping, so malformed TOON from upstream tools surfaces immediately
  • Ships a concrete temp-directory pattern with an EXIT trap for persisting conversions during a task

Limitations

  • Assumes the toon CLI is installed and available; if missing, the skill says to surface the failure rather than silently fall back to raw JSON
  • Handles one JSON value at a time, so NDJSON/JSONL must first be slurped into an array with jq -s
  • TOON is a read-only surface: JSON stays canonical, and the skill explicitly discourages adding a TOON authoring step
When to use it

Use it when a task must inspect a large JSON file, API response, set of database rows, or structured log and reading the raw JSON directly would be expensive in tokens.

When not to use it

Avoid it when the data is small, is not JSON-compatible, or when the task is to modify and write the data rather than read it.

Security analysis

Caution
Quality score88/100

The skill is not malicious: it does not exfiltrate secrets, use obfuscated payloads, or disable safety. However, it instructs the agent to run shell commands, access the network, and perform scoped rm -rf cleanup, so it warrants caution rather than safe.

Findings
  • •Uses shell commands and network access (jq, curl, toon, mktemp, rm) despite no allowed-tools being declared.
  • •Includes curl piped into a converter for remote API data; fetch target should be trusted, though no secrets are sent.
  • •Uses rm -rf for cleanup, scoped to a mktemp directory, but still a destructive command pattern.
  • •Relies on an external toon CLI whose availability and behavior are not verified by the skill.
How this analysis is made

Examples

Inspect a large API response
Fetch https://api.example.com/items and inspect the response with toon instead of reading the raw JSON, then tell me which items have status "failed".
Trim fields before encoding
Use jq to keep only id, name, and status from response.json, pipe it through toon, and summarize the results.
Convert NDJSON logs
Slurp events.ndjson into an array with jq -s and run it through toon so we can inspect the structured logs without blowing up context.

name: toon description: Use the TOON CLI whenever a task inspects JSON-compatible structured data such as JSON files, API responses, database rows, search results, and structured logs. This skill applies before that data enters agent context.

TOON

Purpose

TOON is a temporary, lossless representation of the JSON data model for reducing the tokens required to read structured data.

JSON remains the canonical file and interchange format. TOON is used only as a reading surface for inspection.

Input Normalization

TOON accepts one JSON value rather than newline-delimited JSON (NDJSON/JSONL), so NDJSON is slurped into an array before encoding:

jq -s '.' events.ndjson | toon

Inputs are reduced to the required rows and fields before encoding when the task needs only part of the data:

jq '{items: [.items[] | {id, name, status}]}' response.json | toon

Reading Structured Data

The workflow reads the CLI's TOON output instead of first reading the source JSON into agent context.

toon large-response.json
curl -fsSL https://api.example.com/items | toon

Converted data that must persist during the task is written under a system temporary directory owned by the current command sequence:

scratch="$(mktemp -d -t toon.XXXXXX)"
cleanup() {
  status=$?
  rm -rf "$scratch"
  exit "$status"
}
trap cleanup EXIT
toon large-response.json --stats -o "$scratch/large-response.toon"

The complete source JSON and its TOON representation are not both read unless comparison is required by the task.

Validation

The toon command uses strict decoding by default. Strict decoding validates array counts, indentation, headers, and escaping, so malformed TOON input from upstream tools is surfaced immediately.

--no-strict is not used to accept malformed TOON. Missing CLI availability or conversion failures are surfaced instead of silently falling back.

Required Syntax

Objects use key: value, and nesting uses two-space indentation:

user:
  id: 1
  name: Alice
  active: true

Primitive arrays declare their element count:

tags[3]: rust,cli,macos

Uniform arrays of objects declare the row count and field order once:

users[2]{id,name,active}:
  1,Alice,true
  2,Bob,false

[N] is the array element count. {fields} defines the field order for every following row.

Quote strings that could be interpreted as another type or contain structural characters:

version: "123"
enabled: "true"
note: "hello, world"

These rules matter when reading existing TOON output or interpreting examples. Do not introduce a TOON authoring step when the task can stay in JSON.

From the file to a first result

The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.

View the source and supporting files

This download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.

1. Download the file

Download SKILL.md

From the root of a test project, save the file as .skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.

Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.

mkdir -p '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/toon-json-data-inspection-tool-mrwakg7k/download' --output '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md.part' && mv '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md.part' '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md'

2. Check prerequisites

Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.

No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.

The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.

3. Load the skill

After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.

Claude Code
test -f '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md' && test -s '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/toon' && cp -R '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k' '.claude/skills/toon'

Cursor
test -f '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md' && test -s '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md' && mkdir -p '.cursor/skills' && test ! -e '.cursor/skills/toon' && cp -R '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k' '.cursor/skills/toon'

Codex
test -f '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md' && test -s '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md' && mkdir -p '.agents/skills' && test ! -e '.agents/skills/toon' && cp -R '.skill-downloads/toon-json-data-inspection-tool-mrwakg7k' '.agents/skills/toon'

OpenSkills

OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.

Official OpenSkills installation
openskills install './.skill-downloads/toon-json-data-inspection-tool-mrwakg7k'

openskills read 'toon-json-data-inspection-tool-mrwakg7k'

Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.

For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.

openskills sync

Manual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.

4. Try a first task

Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:

Read .skill-downloads/toon-json-data-inspection-tool-mrwakg7k/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.

Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.

Related skills