Scope of `consequentialHint` for high-risk read tools
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 35/100
- Issue 类型
- 功能
- 描述清晰度
- 基本清楚
- 活跃度
- 活跃
- 技术栈
- html
- 领域
- api
调研方向
Start by reading the WebMCP spec definition of consequentialHint and the linked pull request 217, then review the six comments for agreement on the scope. Decide whether to extend the hint to cover high-risk data exposure or define a separate sensitive-data hint; done means the spec clearly distinguishes the relevant consent risks and defines the chosen hint.
由索引模型根据 Issue 内容生成。
描述
The current WebMCP spec defines consequentialHint (introduced in https://github.com/webmachinelearning/webmcp/pull/217) as:
"If true, indicates that executing the tool will result in consequential actions that are significant, real-world, or non-reversible, ex: booking a flight, transferring money."
This definition focuses on mutative operations (write actions) but leaves high-risk read-only operations ambiguous.
Many tools do not alter state but retrieve highly sensitive, restricted, or regulated user data. Letting an agent execute these silently without user confirmation introduces severe privacy and security risks.
| Tool Example | Risk Category | Potential Impact |
|---|---|---|
readUserAddress |
Location Privacy | Exposes the user's physical, real-world location. |
readBankAccountBalance |
Financial Privacy | Reveals net worth, enabling targeted financial exploits. |
readApiKey |
Security Credentials | Exposes secrets that grant permanent external write access. |
readMedicalHistory |
Regulated Data | Exposes highly sensitive health records (e.g., HIPAA). |
I'd be in favor of updating consequentialHint to explicitly include high-risk data exposure with something like:
*"If true, indicates that executing the tool or disclosing its results will lead to significant, real-world, non-reversible, or high-privacy-risk consequences (e.g., booking a flight, transferring money, or exposing credentials and medical records)."
OR we could also introduce a new hint (e.g., sensitiveDataHint) to keep consequentialHint strictly reserved for side-effect-producing mutations, and introduce a separate, parallel hint specifically for data access risks. This would allows agents to differentiate their consent UI (e.g., "This tool will execute an action" vs. "This tool will read sensitive information").
@johannhof
- 主要语言
- Bikeshed
- 星标
- 4.7k
- 派生
- 328
- 平均合并
- 2 天 15 小时
- 30 天内合并 PR
- 16
环境准备
这个项目没有提供开发容器、Dockerfile 或贡献指南，环境需要你自己搭建：先看它的 README，通用步骤见我们的新手贡献指南。
从这里开始
- 先读完整个 Issue，再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库，在一个分支上完成修改。
- 提交 Pull Request，并在描述里引用这个 Issue 编号。
webmachinelearning/webmcp 的其他 Issue
-
难度 1/5 1-3 小时 新手友好度 72/100
webmachinelearning/webmcp#360 · 1 条评论 ·
维护者通常 1 天内回复
-
Say how an agent should weigh annotations that the page wrote about itself可能已有人在做 关联的 PR 仍在进行中或已合并。 未关闭
难度 2/5 1-3 小时 新手友好度 72/100
webmachinelearning/webmcp#353 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 67/100
webmachinelearning/webmcp#343 · 1 个 reaction ·
维护者通常 1 天内回复
-
Agenda+ i18n-tracker
难度 2/5 1-3 小时 新手友好度 75/100
webmachinelearning/webmcp#273 · 3 条评论 ·
维护者通常 1 天内回复
-
a11y-tracker Agenda+
难度 2/5 1-3 小时 新手友好度 67/100
webmachinelearning/webmcp#272 · 4 条评论 · 2 个 reaction ·
维护者通常 1 天内回复
查看 webmachinelearning/webmcp 的全部 Issue
相似的 Issue
-
难度 2/5 1-3 小时 新手友好度 85/100
gramaziokohler/compas_timber#870 ·
-
难度 1/5 1 小时以内 新手友好度 88/100
维护者通常 3 天内回复
-
bug
难度 2/5 1-3 小时 新手友好度 75/100
eugenioenko/autentico#548 · 1 条评论 ·
维护者通常 1 天内回复
-
sidecar: --moriio-parallel-dispatch omits remote_request_id, so a vLLM NixlPushConnector decode dies with KeyError可能已有人在做 @xintin 今天认领。 未关闭kind/bug needs-triage
难度 2/5 1-3 小时 新手友好度 78/100
llm-d/llm-d-router#3607 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 75/100
swiftlang/indexstore-db#306 ·