I 'm very interested in RBAC as it's convenient to grant user access to different projects and operations.
- For a single user, RBAC is useless at the first glance. However, multi-agents could also be integrated into RBAC.
- Sub-agents could derive roles from users to access certain projects within certain operations, to prevent them doing any shit unexpected.
- For instance, a research agent has only read access, so it won't automatically generate a research summary that pollutes the environment.
- Another example is that we could grant sub agent write access only to certain folders, to prevent it modify any other places that are irrelavant to the mission.
- For a team that use agents on a project, they need to share not only the project itself, but also the available tools, the Agent memories, and the raw documents.
It would be easy of use, if this project would offer some modules as agent memory plugins, to allow a team to share the memories of their agents and with rigid access control.
- Some experts could upload their memories on certain modules, interfaces, etc,
- Some TLs could act as the arbitrators to deal with memory conflicts;
- Some inexperienced members are just users that could just read the memories.
It's understandable that the RBAC system and memory system are 2 independant modules that should not appear in one repos, but it would be great if we can have 2 projects that can be used together easily.
I am currently working on the shared Agent Memory system and how the memories are distributed on each agent container. This project is enlightening and give me some ideas regarding how RBAC module and shared-memory module could be integrated.
I 'm very interested in RBAC as it's convenient to grant user access to different projects and operations.
It would be easy of use, if this project would offer some modules as agent memory plugins, to allow a team to share the memories of their agents and with rigid access control.
It's understandable that the RBAC system and memory system are 2 independant modules that should not appear in one repos, but it would be great if we can have 2 projects that can be used together easily.
I am currently working on the shared Agent Memory system and how the memories are distributed on each agent container. This project is enlightening and give me some ideas regarding how RBAC module and shared-memory module could be integrated.