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 onescience-ai/OneSkills --skill onescience-parallelgit clone --depth 1 https://github.com/onescience-ai/OneSkillsWrote 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/onescience-ai/oneskills/onescience-parallel)<a href="https://agentmods.dev/skills/onescience-ai/oneskills/onescience-parallel"><img src="https://agentmods.dev/badge/skills/onescience-ai/oneskills/onescience-parallel/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/onescience-ai/oneskills/onescience-parallel"><img src="https://agentmods.dev/badge/skills/onescience-ai/oneskills/onescience-parallel.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 | $0.00210 | $0.10681 |
| Opus 5.5 | $0.00084 | $0.04272 |
| Sonnet 5.5 | $0.00042 | $0.02136 |
| Haiku 4.5 | $0.00021 | $0.01068 |
Grade A, and why
onescience-parallel 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 6d 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 — 1,056 lines — stays where its author put it; the contents beside it link to each section on GitHub.
输入获取方式
本技能支持两种输入方式:
- 上下文 handoff(默认):从调用方传入的
step_handoff获取任务信息。 - 文件 handoff(autonomous_mode):从
.onescience/handoff/step_{step_id}.yaml读取任务信息。执行后,将结果写入.onescience/handoff/step_{step_id}_result.yaml。
启动时优先检查 .onescience/handoff/ 目录是否存在对应的交接文件;若存在则使用文件模式,否则使用上下文模式。
文件交接格式参见 skills/onescience-orchestrator/references/file_handoff_contract.md。
Pipeline Parallel 改造 Skill
重要原则(必读)
- 复用 OneScience 模块:所有基础模块必须使用
onescience已有实现,禁止重复实现 - 先读后写:开始前必须阅读
context.md和architecture.md - 参考已有实现:必须参考
examples/earth/pangu_weather_distributed/的 pangu 实现 - 保持参数一致性:进行并行改造时,各 Stage 类的
__init__参数签名和内部初始化逻辑应尽可能与原模型保持完全一致(如config、mask等参数的处 理),避免随意更改参数名或逻辑。 - 严禁循环导入:在创建 Distributed 模块(如
{StyleName}DistributedFuser)时, 禁止在模块内部导入OneFuser、OneTransformer等顶层包装类 ,因为这些包装类通常已经导入了你的 Distributed 模块,会导致ImportError。应直接导入具体的子模块类(如from .{stylename}distributedlocalsiefuser im port {StyleName}DistributedLocalSIEFuser)。
改造三步流程
步骤 1:模型拆分 (PP) → 步骤 2: TP 并行模块 → 步骤 3:训练接口对接
步骤 1:模型拆分(Pipeline Parallel)
1.1 切分策略
按前向执行顺序找计算串行边界,切点满足:
- 数据依赖最少(只有一个 tensor 出口)
- 跨 stage 传输的中间 tensor 尽量小
- 各 stage 计算量尽量均衡
典型 4-stage 切分(U-Net/Encoder-Decoder 结构):
| Stage | 内容 | 说明 |
|---|---|---|
| 0 | Embedding + Encoder 前半 | 首阶段,产生 skip connection |
| 1 | Downsample + 中间层 | 下采样后的计算密集区 |
| 2 | 解码层 + Upsample | 解码器前半段 |
| 3 | Decoder 后半 + Recovery | 消费 skip,产出最终结果 |
1.2 每个 Stage 的必要属性
class MyModel_stageN(Module):
def __init__(self, original_arg1, original_arg2, ..., megatron_config=None):
"""
参数签名应尽可能与原模型保持一致。
如果原模型第一个参数是 config (yaml配置 ),则保持不变;
额外传入的 Megatron 核心配置建议命名为 megatron_config 避免冲突。
"""
super().__init__(meta=MetaData())
# ① 必须有这三个属性,均设为 None
self.pre_process = None
self.share_embeddings_and_output_weights = None
self.input_tensor = None # Stage 0 也要有
# ② 必须初始化 config(Megatron get_model_config() 需要)
# 使用传入的 megatron_config,若无则从 args 获取
if megatron_config is None:
args = get_args()
megatron_config = core_transformer_config_from_args(args)
self.config = megatron_config
# ③ 保持原模型的初始化逻辑
self.arg1 = original_arg1
# ... 原模型的参数处理 ...
def set_input_tensor(self, input_tensor):
"""Megatron pipeline 调度钩子,所有 Stage 都必须实现 """
self.input_tensor = input_tensor
What ships with it
3 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.
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.
- 6d ago Changed · +43 lines 6e3d2d1af2d2
- 2mo ago First seen · 1,013 lines · 210 tokens per session scan A 403631e0dec4
onescience-parallel is a skill published in the GitHub repository onescience-ai/OneSkills (24 stars, last pushed yesterday), licensed MIT. It adds 210 tokens to every session and 10,681 once invoked, about $0.0008 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 2026-08-30.
Other skills, from other repositories
graph-retrieval
Exposes graph-based retrieval as a tool capability via querygraph. Reads normalized graph store files, builds a query-relevant subgraph, and returns LLM-friendly semantic triples with replayable evidence metadata.
knowledge-layer
High-level deployment wrapper over RepoBrain core with graph-first knowledge injection and all-file support. Exposes refreshfilesystem and askfilesystem for building and querying the knowledge graph.
n8n-agents
Design n8n AI agents the right way. Use when building or editing any @n8n/n8n-nodes-langchain. AI node — an AI Agent, LLM chain, Text Classifier, or Information Extractor — and whenever the user mentions AI agents, LLM with tools, tool calling, $fromAI, system prompts, agent memory, sessionId, structured/JSON output…
bioprobench
Score an LLM's biological-protocol reasoning on the BioProBench benchmark: protocol QA, step ordering, error detection, protocol generation, and LLM-judged error reasoning; or generate the responses.
td-api-reference
MUST READ before writing TD Python via executepython, setdatcontent, or editdatcontent, and before any heavy build. Parameters, storage, operators, referencing, threading, cook model, heavy-build safety caps.
haiku-rag
Search, read and compute over the user's haiku.rag knowledge base through the haiku-rag MCP tools. Use whenever a request could be answered from the user's ingested documents, when asked to find, look up, check or cite something in their documents or knowledge base, or when the question is about the user's own…