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-knowledge-harvestergit 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-knowledge-harvester)<a href="https://agentmods.dev/skills/onescience-ai/oneskills/onescience-knowledge-harvester"><img src="https://agentmods.dev/badge/skills/onescience-ai/oneskills/onescience-knowledge-harvester/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-knowledge-harvester"><img src="https://agentmods.dev/badge/skills/onescience-ai/oneskills/onescience-knowledge-harvester.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00198 | $0.08314 |
| Opus 5.5 | $0.00079 | $0.03326 |
| Sonnet 5.5 | $0.00040 | $0.01663 |
| Haiku 4.5 | $0.00020 | $0.00831 |
Grade A, and why
onescience-knowledge-harvester 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 25d 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 — 454 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OneScience Knowledge Harvester
你是 OneScience 的知识缺口自动填补执行技能。你的职责是把用户描述的知识缺口转换成可被 onescience-primitives 召回的知识卡片。
核心原则：全程使用智能体原生能力（WebSearch、WebFetch、LLM 推理、Write），不依赖外部流水线或额外 API Key。
证据主体：论文证据（full-text/abstract）是卡片价值的主体；开源权威文档与用户自有数据是低优先级补充通道，只用于论文通道覆盖不到的残余缺口，且必须通过权威性与溯源门禁（见 Step 3.4–3.6）。
触发场景
使用本技能处理以下请求：
- "缺少藏红花基因衍生推理的知识"
- "帮我补充关于 X 的知识"
- "填补 X 领域的知识缺口"
- "自动采集 X 相关的论文知识"
- "我需要 X 方面的知识卡片"
- 任何描述知识缺口并要求自动填补的请求
不要在以下情况下自动触发：
- 用户只要求检索已有知识（那是 onescience-primitives 的职责）
- 用户只要求实时文献推理（那是 onescience-live-literature 的职责）
- 用户只要求沉淀交互经验（那是 onescience-knowledge-capture 的职责）
核心边界
- 本技能是知识采集的执行层，使用智能体原生能力完成全流程。
- 生成的卡片必须符合 onescience-primitives 的格式规范（metadata.json + knowledge.md）。
- 卡片核心内容必须基于实际检索到的论文证据，不得编造；补充通道（开源文档/用户数据）证据必须有权威出处或用户背书，见 Step 3.4–3.6。
- 不修改 onescience-primitives 的 SKILL.md；对已有卡片只允许按 Step 4.5/Step 5 的三态入库（new/modify/merge），且必须通过单调性校验，不得删除内容或卡片目录。
- 生成的卡片按 Step 1 判定的
card_type落盘到对应类目skills/onescience-primitives/assets/<domain>/<category>/<name>/；不得默认把所有卡都写到workflow-planning/——该类目仅承载跨领域的通用方法论卡，且被 onescience-primitives 契约禁止用于回答具体科学问题（详见 Step 1 类目选择表与「禁止事项」）。
执行流程
Step 1 解析知识缺口
从用户输入中提取：
- gap_description：知识缺口的自然语言描述（如"藏红花基因衍生推理"）
- domain：领域（bio/matchem/climate/cfd/general），从描述中自动检测
- search_queries：2~4 组英文检索词（含同义词、物种学名、基因/酶/通路名、方法名）
- card_identity：卡片身份层级。内部先做一步抽象推理：这个缺口来自哪个具体场景 → 该场景属于哪一类通用问题（用一句自然语言描述，如"面向目标气体，选择或设计满足容量、选择性与再生能耗要求的捕集材料"）→ 该问题的解法能否迁移到同类其他体系。能迁移 → 卡片身份 = 通用问题类（生成通用类卡）；缺口只是单一实例事实、无可迁移方法层 → 不强行升维，卡片身份 = 实例级。这条推理链只用于确定卡片身份，严禁以任何形式出现在卡片产物中（见 Step 4 通用类卡撰写规范）。
- card_type + category（硬门禁，决定落盘类目与 metadata.type）：按缺口性质从下表选择唯一类型，不得默认落到
workflow-planning。判定证据（缺口关键词/归因报告 issues 类型/所需产物形态）必须写入 execution_result.observation.type_rationale。
card_type 选择表（按优先级从上到下匹配，命中即停）：
| 缺口性质 | card_type | category（落盘目录） | 命中示例 |
|---|---|---|---|
| 具体领域的任务级 know-how：某类计算/实验/分析任务的做法、参数、验收 | task |
<domain>/tasks/ |
"CO2 在 MOF 上的 GCMC 吸附等温线计算"、"RNA-seq 差异表达分析" |
| 具体领域的场景级 know-how：一个完整科研场景的问题设定、判据、边界 | scenario |
<domain>/scenario/ |
"面对含湿烟道气的 CO2 捕集材料筛选"、"MW 级电池热失控预警" |
| 具体领域的工作流级 know-how：多个 task 的编排序列、依赖、数据流 | workflow |
<domain>/workflow/ |
"DFT→机器学习势→MD 的跨尺度工作流" |
| 具体的资源实体：模型、数据集、工具、组件、数据库、服务、应用、可视化、契约、输出格式 | model / dataset / tool / component / database / service / application / visualization / contract / output-format |
对应同名目录 <domain>/<category>/ |
"MACE-MP-0 势"、"Materials Project API"、"pymatgen 结构分析工具" |
| 跨领域的通用方法论：科研规划、通用方法族、通用建模流程、通用筛选策略 | workflow-planning |
<domain>/workflow-planning/ |
"面向任意目标气体的捕集材料筛选方法论"、"通用 CLI 故障分类" |
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.
- 25d ago Changed · +19 lines 410431d55c1e
- 29d ago Changed · +2 lines 7a41d2441e38
- 1mo ago Changed · +119 lines · +49 tokens per session df76fe12675b
- 1mo ago First seen · 314 lines · 149 tokens per session scan A a3ce130d5171
onescience-knowledge-harvester is a skill published in the GitHub repository onescience-ai/OneSkills (23 stars, last pushed yesterday), licensed MIT. It adds 198 tokens to every session and 8,314 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 2026-09-11.
Other skills, from other repositories
contorium-workspace-memory
Use Contorium workspace memory layers (focus, session events, Git, export) when resuming AI work or avoiding repeated project explanation.
locus-memory
Use when the user needs persistent project memory across sessions. Triggers on: "remember this", "what did we decide about...", "show project structure", "what changed recently", "search memory". Do NOT trigger for: ephemeral questions, one-off lookups, file reads that don't need memory.
remember
Save a project decision or context to Locus semantic memory. Use when the user says "remember this", "save this decision", or wants to store architectural context for future sessions.
compact
Clean up old episodic memory entries to free space. Keeps recent sessions intact.
forget
Delete memory entries from Locus matching a search query. Use when the user says "forget this", "delete memory", or "remove that decision".
memory-audit
Audit Locus stored data — shows inventory of structural map, semantic memories, episodic entries, hook captures, DB size, and security status. Use when the user wants to know what data Locus has stored.