Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add soulcodex/agentic --skill json-to-toongit clone --depth 1 https://github.com/soulcodex/agenticWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/soulcodex/agentic/json-to-toon)<a href="https://agentmods.dev/skills/soulcodex/agentic/json-to-toon"><img src="https://agentmods.dev/badge/skills/soulcodex/agentic/json-to-toon/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/soulcodex/agentic/json-to-toon"><img src="https://agentmods.dev/badge/skills/soulcodex/agentic/json-to-toon.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00088 | $0.01377 |
| Opus 5.5 | $0.00035 | $0.00551 |
| Sonnet 5.5 | $0.00018 | $0.00275 |
| Haiku 4.5 | $0.00009 | $0.00138 |
Grade A, and why
json-to-toon scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
JSON to TOON Transformation Skill
Convert JSON input into TOON (Token-Oriented Object Notation) to achieve 30–70% token reduction, making it ideal for use in LLM prompts, RAG pipelines, and any context-window-sensitive application.
Step 1 — Load the Specification
Read toon-spec-reference.md (in this skill's directory). It contains the
complete encoding rules you must apply throughout this skill.
Step 2 — Assess the Input
Inspect the JSON the user has provided:
- Identify the root type: is it an object, array of uniform objects, array of mixed objects, or a primitive array?
- Scan for high-repetition patterns: arrays of objects with identical keys are prime candidates for tabular format (biggest token savings).
- Flag any values that require quoting: strings containing
,,:,[,{,-at start, leading/trailing spaces, or values that look like booleans/numbers/null. - Note depth: deeply nested structures may benefit most from TOON's indentation-based nesting (no closing braces).
Output a brief one-line assessment: e.g.
"Root object with 3 tabular arrays and 2 nested objects — expect ~55% token reduction."
Step 3 — Choose Array Format for Each Array
For every array in the input, apply this decision tree:
Is every element a primitive (string, number, boolean, null)?
YES → Inline format: key[N]: v1,v2,v3
Are ALL elements objects AND do they share identical top-level keys AND
are all their values primitives (no nested objects or arrays)?
YES → Tabular format: key[N]{f1,f2,f3}:
v1,v2,v3
v4,v5,v6
Otherwise (different keys, nested values, mixed types)
→ List format: key[N]:
- field: value
field2: value2
Use the field order of the first object for all tabular rows. Never re-order fields.
Step 4 — Encode the Full Structure
Walk the JSON top-down and apply the rules from toon-spec-reference.md:
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 159 lines · 88 tokens per session scan A 59ac49be61b2
json-to-toon is a skill published in the GitHub repository soulcodex/agentic (10 stars, last pushed 1mo ago), licensed MIT. It adds 88 tokens to every session and 1,377 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
qdrant
Manage Qdrant vector database via REST API. Use when the user asks to create or delete collections, upsert or search vectors, inspect points, filter by payload fields, manage snapshots, check cluster status, or debug semantic search issues. Covers collection CRUD, point upsert/search/scroll/count, payload indexes…
hatch3r-cli-jq
JSON processor and query language. Use when shaping JSON streams via jq-syntax filters and select expressions; invoke jq. Reads stdin and emits stdout; integrates seamlessly into shell pipelines.
datoon
Smart TOON conversion workflow for structured data in Claude Code. Converts JSON-like payloads to TOON only when structure is suitable and token savings are meaningful.
Training
Guides fine-tuning and post-training work with explicit data, objective, hardware, and rollback assumptions.
evaluating-llms-harness
Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.
serving-llms-vllm
Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.