dataflow-pattern-math-reasoning-sft

dataflow-pattern-math-reasoning-sft is a skill for Claude Code, Codex from OpenDCAI/Dataflow-LoopAI. It costs 37 tokens per session (336 once invoked), scanned A, original, Apache-2.0.

A reusable pattern for creating training examples from short maths questions with checkable answers. SFT means supervised fine-tuning, a way to train a model from example inputs and desired outputs.

In plain words
What is it for?
Turning question-and-answer or problem-and-final-answer data into validated instruction, reasoning, and answer records for benchmark-aligned maths training.
Why use it?
It keeps the original sample identity, hides the known answer during reasoning generation, checks the generated result against that answer, and requires complete output fields.

Skill for Claude CodeCodex

Install

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.

agentmods
npx agentmods add skills/opendcai/dataflow-loopai/complete-math-reasoning-sft-skill
Any agent
npx skills add OpenDCAI/Dataflow-LoopAI --skill complete-math-reasoning-sft-skill
Clone the repo
git clone --depth 1 https://github.com/OpenDCAI/Dataflow-LoopAI

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for dataflow-pattern-math-reasoning-sft

README.md
[![agentmods](https://agentmods.dev/badge/skills/opendcai/dataflow-loopai/complete-math-reasoning-sft-skill.svg)](https://agentmods.dev/skills/opendcai/dataflow-loopai/complete-math-reasoning-sft-skill)
Your own site
<a href="https://agentmods.dev/skills/opendcai/dataflow-loopai/complete-math-reasoning-sft-skill"><img src="https://agentmods.dev/badge/skills/opendcai/dataflow-loopai/complete-math-reasoning-sft-skill.svg" alt="Measured on agentmods" height="20"></a>
Per session 37 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 336 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00037 $0.00336
Opus 5 $0.00018 $0.00168
Sonnet 5 $0.00007 $0.00067
Haiku 4.5 $0.00004 $0.00034

Measured 5d ago against content hash 1e7c56d4e300, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

dataflow-pattern-math-reasoning-sft 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 5d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (examples/gsm8k-reasoning-v1/pipeline.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

loopai/agents/Obtainer/datamixer/assets/curating-dataflow-pipeline-skills/example/complete-math-reasoning-sft-skill/SKILL.md · 33 lines

What it actually says

Math Reasoning SFT Pattern

Use this pattern for short, verifiable math problems whose training output requires a standalone instruction, regenerated reasoning, and a final answer compatible with strict benchmark evaluation. Do not use it for proof-only tasks, open-ended tutoring, or records without a trustworthy answer signal.

The recommended case is gsm8k-reasoning-v1. It accepts question/answer and problem/final_answer schema branches, normalizes them to canonical fields, regenerates reasoning without exposing the gold answer to the generator, validates the derived final answer against gold, and applies an LLM quality evaluation after deterministic correctness checks.

Preserve these invariants when adapting the case:

  • sample_id and source dataset provenance survive every stage.
  • The reasoning generator sees the problem but not the gold answer.
  • Generated reasoning is accepted only when its parsed conclusion matches gold.
  • Quality thresholds are justified by the trial score distribution.
  • Output always provides nonempty instruction, reasoning, and output.
  • Benchmark answer formatting is handled explicitly rather than left to chance.

For a new schema or benchmark, retain the staged validation pattern but adapt normalization, correctness parsing, difficulty coverage, and final formatting to the actual task.

Files

What ships with it

6 files 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.

Changes

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.

  1. 5d ago First seen · 33 lines · 37 tokens per session scan A 1e7c56d4e300

Subscribe to this mod's changes

dataflow-pattern-math-reasoning-sft is a skill published in the GitHub repository OpenDCAI/Dataflow-LoopAI (24 stars, last pushed 3d ago), licensed Apache-2.0. It adds 37 tokens to every session and 336 once invoked, about $0.0002 per session on Opus 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 2026-08-30.

Related

Other skills, from other repositories

method-design

Use this skill whenever the user wants to formalize a network architecture and derive theoretical components from a research idea. Triggers include: 'method design', 'design method', 'network architecture', 'formula derivation', 'method-design', 'theoretical framework', 'derive equations', 'compare alternatives'…

CUHK-AIM-Group/NeuroClaw · 163 tokens

llm-tuning-patterns

Evidence-based patterns for configuring LLM parameters, based on APOLLO and Godel-Prover research.

parcadei/Continuous-Claude-v3 · 12 tokens

seo-llm

Use when optimizing content for LLM-powered search engines (ChatGPT, Perplexity, Gemini, Claude, Bing AI, Qwen), implementing RAG optimization, prompt engineering for search visibility, semantic SEO, and ensuring content ranks highly in AI-driven search results. Includes techniques for ChatGPT SEO, Perplexity…

Omar-Obando/qwen-orchestrator · 79 tokens

llm-integrations

Use when integrating LLM providers (OpenAI, DeepSeek, OpenRouter, Anthropic, Google), configuring API keys, optimizing costs, implementing rate limiting, and managing LLM usage across projects. Includes best practices for cost optimization and API management. Based on OpenAI, Anthropic, Google, and other LLM provider…

Omar-Obando/qwen-orchestrator · 73 tokens

langchain

Use when building LLM applications with LangChain, implementing chains, agents, tools, memory, prompts, and retrieval systems. Includes best practices for prompt engineering, tool integration, and agent development. Based on LangChain/LangGraph official documentation and agent development best practices.

Omar-Obando/qwen-orchestrator · 57 tokens

enhance-prompt

Transforms vague UI ideas into polished, Stitch-optimized prompts. Enhances specificity, adds UI/UX keywords, injects design system context, and structures output for better generation results.

diodeme/Gold-Band · 41 tokens