osFoundry / Platform
Guards
在执行前拦截危险的工具调用——破坏性 shell、违反策略的 URL、PII 写入。可按工作区配置。
简介
Guards is an osFoundry feature for block dangerous tool calls before they run — destructive shell, off-policy URLs, PII writes. Configurable per workspace.. It keeps this work in the same workspace as your conversations, AI agents, files, and projects, so people can move from information to action without switching tools.
Use Guards when you need a practical way to organize, monitor, configure, or automate this part of your workflow. The feature is designed to work with the rest of your osFoundry workspace, making it easier to share context, collaborate with your team, and build repeatable processes.
主要功能
- Use Guards to block dangerous tool calls before they run — destructive shell, off-policy URLs, PII writes. Configurable per workspace.
- Keep Guards connected to related workspace tools and AI agents.
- Review and manage Guards from one focused workspace.
使用场景
- Set up Guards for a new project or team workflow.
- Use Guards while work is in progress and review the result when it is complete.
- Connect Guards with related osFoundry tools, data, and AI agents.