TOON

Optimize token usage by adopting the compact TOON format for data storage and context.

Security Scan

This is an instruction-only skill that defines a compact data format (TOON) and how agents should parse/write it; its requirements and instructions are coherent with that purpose and it does not request credentials, installs, or external endpoints.

Purpose & Capability

The name/description (TOON format for token efficiency) matches the content: a syntax spec, examples, and agent usage guidelines. No unrelated binaries, environment variables, or install actions are requested.

Instruction Scope

SKILL.md and README stay within the stated purpose (parsing/generating/storing TOON). They recommend using .toon files for logs, schedules, and context compression; this is expected but worth noting because it encourages the agent to persist potentially sensitive data in a new file format. README also shows copying/unzipping into /a0/usr/skills/, which is a normal agent install path but implies write access to the agent skills directory.

Install Mechanism

No install spec or code files are included (instruction-only). That is the lowest-risk install profile — nothing is downloaded or executed by the skill itself.

Credentials

The skill does not request any environment variables, credentials, or config paths. There are no disproportionate or unexplained secret requests.

Persistence & Privilege

Flags show normal behavior (always: false, user-invocable true). The skill does not request permanent presence or elevated privileges beyond recommending a location in the agent's skills directory.

Assessment

This skill is a coherent format specification and agent guidance for using TOON; it does not contact external services or request credentials. Before installing, consider: 1) The skill encourages storing logs/summaries in .toon files — ensure you are comfortable with the agent writing potentially sensitive data to disk and set appropriate file permissions and retention policies. 2) The claimed ~40% token savings is a guideline — test conversion on representative data to confirm benefits. 3) If you integrate TOON with other tooling, validate converters/parsers for correctness (lossless conversion to/from JSON). Overall the artifact appears consistent and low-risk, but treat any persisted logs as you would other sensitive agent data.

TOON Adoption Skill

Description

Efficiently parse, generate, and store data using Token-Oriented Object Notation (TOON). TOON is designed for LLMs as a lossless, drop-in representation of JSON data that reduces token usage by approximately 40% through indentation, minimal quoting, and tabular layouts.

TOON Syntax Rules

1. Indentation-based Nesting

• Use 2-space indentation to define hierarchy, similar to YAML.
• Avoid curly braces {} and square brackets [] for nesting unless defining a tabular schema.

2. Minimal Quoting

• Keys and values do not require quotes unless they contain commas or significant leading/trailing whitespace.

3. Explicit Array Lengths

• Declare the number of items in an array using the [N] notation (e.g., friends[3]).

4. Tabular Arrays (Arrays of Objects)

• For uniform arrays of objects, use the format: key[N]{field1,field2,...}:.
• List the field names once in the header, followed by rows of values separated by commas.

5. Encoding

• Always use UTF-8.

Guidelines for Agent Use

1. Storage: Prioritize .toon files for logs, schedules, and long-term data tracking.
2. Context Compression: When summarizing history or notes, use TOON to fit more information into the context window.
3. Parsing: Interpret indentation as nested object levels and CSV rows as objects matching the header schema.

Example Document

metadata:
  name: Sample Configuration
  format: TOON
  efficiency_gain: 0.4
goals[3]{id,title,category}:
  1,Lose weight,health
  2,Increase self-esteem,personal
  3,Get out of loneliness,social