使用方法
- 复制安装命令，在自己的智能体环境安装前先检查源码包。
- 从原始文档核对所需工具、凭据和支持的宿主。
- 说明输入材料与期望交付物，并对照来源检查输出。
npx skills add https://github.com/yogsoth-ai/de-anthropocentric-research-engine --skill attach-evidence-to-relation开始前需了解
- 安装不包含模型订阅、API 额度或访问文件的权限。
- 仓库星数和安装量反映活跃情况，不等于独立质量或安全测试。
仓库原始内容
按源仓库内容呈现，保留标题、案例、代码、表格、链接以及原文引用的演示图片。
attach-evidence-to-relation
Purpose
Attach source evidence to a typed relation between concepts, claims, or entities with directness and independence metadata.
Input contract
yaml
required: [relation, evidence_records, relation_schema]
optional: [quality_rubric, independence_keys, confidence_prior]
constraints: [evidence must support the relation type and direction claimed]Procedure
- Identify the relation type, endpoints, direction, and boundary conditions.
- Match each evidence record to the relation and classify directness.
- Assess quality, independence, consistency, and alternative interpretations.
- Emit an evidence-linked relation with confidence rationale.
If attached evidence supports incompatible relation states under the same scope, consider detect-contradiction as the next tactic.
Output contract
yaml
produces: [evidence_relation, support_records, directness_assessment, independence_assessment, confidence_rationale]
delta_fields: [findings, evidence_updates, uncertainties, decisions, open_questions]Quality gates
- Evidence supports the typed relation rather than only its endpoints.
- Dependent sources are not counted as independent.
Failure and counterexamples
Do not attach a citation merely because it mentions both entities.
Provenance map
resolved: attach-evidence-to-relation
