# Language-based security

**URL:** <https://community.anytype.io/t/language-based-security/19381>\
**Category:** Talk\
**Tags:** Design-or-UI, Community, Security-and-Privacy, Plugins\
**Created:** [April 2, 2024, 6:57am UTC](https://community.anytype.io/t/language-based-security/19381 "2024-04-02T06:57:46Z")\
**Posts on this page:** 1\
**Showing post:** 2

<div class="post-metadata">

**Author:** ![micklat](https://community.anytype.io/user_avatar/community.anytype.io/micklat/32/15411_2.png) [@micklat](https://community.anytype.io/u/micklat)\
**Post date:** [April 2, 2024, 11:31am UTC](https://community.anytype.io/t/language-based-security/19381/2 "2024-04-02T11:31:19Z")

</div>

A simple implementation of [Derived Relations](https://community.anytype.io/t/derived-relations-formulas/1554) and [in-line calculations](https://community.anytype.io/t/add-in-line-calculations-the-power-of-excel-but-in-free-text-like-patera/1709) does not require an advanced type system. It’s probably enough that the formula language be pure (no side-effects) and that formulas cannot reference data from other spaces (assuming that data permissions are set at the space, and not the object level).

As for plugins (as requested [here](https://community.anytype.io/t/how-will-should-the-plugins-system-work-for-developers-and-users/2485) [here](https://community.anytype.io/t/what-are-your-top-3-requests-for-2024/13279/9) and [here](https://community.anytype.io/t/plugin-architecture-discussion/6790)), this is where things get more involved, because a plug-in system should facilitate dangerous things such as plug-ins writing into objects (possibly overwriting user data), and communicating with the outside world (which can leak sensitive data if allowed too permissively).

So I suppose we can start with the relatively safe formulas and proceed in parallel to do a careful research and design phase for establishing a security framework for plug-ins.

---

_[View the full topic](https://community.anytype.io/t/language-based-security/19381)._
