Format structured data into TOON v2.0 with configurable array types and delimiters.
npx skills add https://github.com/gounthar/bazel-riscv --skill toon-formatter
Please help me install this Agent Skill. Skill: toon-formatter Source: https://github.com/gounthar/bazel-riscv/tree/main/.claude/skills/toon-formatter Command: npx skills add https://github.com/gounthar/bazel-riscv --skill toon-formatter
Aggressively formats large structured datasets into TOON v2.0 to save tokens and enable efficient prompts.
Use the tool to encode a 10-item dataset into TOON with a tab delimiter.
High-intent search queries and answers about installing and using this skill.
Token usage for structured data drops significantly with TOON v2.0 formatting, which compresses arrays, logs, API responses, and database results by applying aggressive encoding with configurable delimiters and key folding for nested objects.
Use TOON formatting when your dataset contains arrays with at least five items and sixty percent field uniformity—such as tables, events, transactions, or analytics records—especially in RAG pipelines, tool calls, and agent data exchanges where token efficiency matters.
TOON v2.0 supports three array types (tabular, inline, expanded) and three delimiters (comma, tab, pipe), enabling you to select the format that best balances readability and compression for your structured output.
Yes, TOON v2.0 includes key folding for nested objects, allowing you to compress hierarchical data structures while maintaining field uniformity across array items.
TOON formatting is designed specifically for API responses, database results, logs, and analytics data—any uniform structured output where compression and encoding efficiency reduce prompt size.
The optional Zig encoder is an additional performance optimization for TOON formatting that further reduces token usage by applying specialized encoding to already-compressed structured data.
modu-aiCompress structured data into TOON format to reduce LLM token usage.
ksciusDesigns LLM prompts with system prompts, few-shot templates, chain-of-thought guidance, and output formats.
pradyumnasagarAutomate inter-skill data transfer and provenance capture via JSON Material Passports.
windyskyEncode and decode data using TOON format for token-efficient transfer.
