dsh-ontology
用于 DeepSeek Harness 的类型化、可推理本体知识库插件,允许智能体定义、断言和查询领域知识。
AI 分析
核心用途是为 AI 智能体提供持久化的结构化领域知识库,支持模式约束和关系推理。适合需要处理复杂业务逻辑、需要智能体维护长期记忆和知识体系的任务。
安装
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:tancheng33/dsh-ontology说明文档
阅读完整 README ↗Configuration
Override any of these in your profile's cordis.patch.yml:
| Key | Default | Meaning |
|---|---|---|
domain | ontology | Storage domain name; one name is one isolated graph. Must match /^[a-z][a-z0-9_]*$/. |
strict | true | Reject facts that violate the vocabulary. false stores them with the violations recorded, for exploratory modelling. |
inference | true | Allow queries to return entailed facts. |
promptSection | true | Contribute the vocabulary summary to the system prompt. |
promptMaxTerms | 60 | Upper bound on terms in that summary. |
promptOrder | 150 | Where the section lands in the assembled prompt. |
defaultLimit / maxLimit | 50 / 500 | Query result sizes. |
maxEntities / maxFacts | 20000 / 100000 | Capacity guards; a write past the bound fails loudly. |
Running two isolated graphs is just two rows:
- insert:
- id: ontology-team
name: dsh-ontology
config: { domain: team_ontology, strict: true, inference: true, promptSection: true,
promptMaxTerms: 60, promptOrder: 150, defaultLimit: 50, maxLimit: 500,
maxEntities: 20000, maxFacts: 100000 }
(A patch replaces a row's whole config, so restate every key.)