Give a channel its own context
Channels can define a working folder, shared instructions, bot roster, and responder rules, so several roles can work within one explicit context.
MausBot
MausBot, published as OpenMausBot, organizes local agent engines as a team of chat contacts. Explore its role-based channels, execution events, and approval flow to design clearer handoffs and reusable team workflows.
Divide a content task into roles with handoffs
Describe a task that needs separate research, review, or operational roles. The request can use MausBot's team model to define shared material and handoff results, while checking service connections and external-action permissions separately from the portable role instructions.
Specify source material, role boundaries, handoff artifacts, and who approves publication.
Copy the request into a new Floatboat conversation and provide the required material.
Download FloatboatYour request for Floatboat
Preview and copy the request. It is not sent automatically from this page.
Reference: MausBot https://github.com/milind-soni/OpenMausBot Read the original documentation and check the current environment and dependencies. Propose any new installation or tool connection before proceeding. My task: Refer to MausBot’s role and approval model to organize a weekly product newsletter. Define researcher, drafter, and reviewer responsibilities, the files each hands over, and the approval before publication. Give a small sample without creating bots or connecting accounts. Expected deliverable: A role and handoff plan with review checkpoints and one sample content package.
About MausBot and OpenMausBot
MausBot is an open-source desktop application maintained in milind-soni/OpenMausBot. Its README also uses OpenMausBot and records the former OpenGrokBot name. It presents agents as chat contacts, with separate roles, model choices, conversations, and optional computer and application connections.
The interface talks to a local harness server that manages agent processes and converts provider events into a common stream. The documented engines include Claude, Codex, and Grok CLIs. Local transcripts and workspace data do not make every operation offline: model requests, connected services, and cloud computers have their own data paths and access terms.
For a Floatboat workflow, the useful design questions are how to divide responsibility, preserve shared context, and ask for approval at the right point. MausBot’s channels and portable team packages provide concrete examples. Studying that design does not establish an installed integration between the two applications.
Coordinate Roles, Context, and Decisions
MausBot connects channel-specific context, engine activity, and portable team instructions through different responsibilities. Compare its conversation and execution layers while retaining the permission checks needed for local computer use, connected services, and tasks that leave the chat.
Normalize engine activity
A local server owns the agent processes and exposes a common event stream. Tool activity and permission requests can be presented without making every engine’s protocol the user interface.
Package a repeatable team
Team files can describe bots, channels, playbooks, and suggested routines. Import review and paused routines separate reusable instructions from live permissions.
Use Cases for MausBot
Research review, recurring operations, and team playbook transfer need shared material without shared credentials. Examine MausBot's role-based approach while distinguishing the draft handoff, service reconnection, and authorization required for actions outside the conversation.
Separate research from review
Research and verification can be assigned to separate roles when both work from traceable source material. Define the shared folder, citation expectations, and evidence required at handoff, then have the review role check claims against those sources. The package should make disagreements and unsupported claims easy to locate.
Prepare a recurring operations brief
A recurring operations brief needs a schedule, approved information sources, and a report format. Design the routine around preparing a draft and handing it to review. Sending messages or changing external records remains a separate action, with the required connection and authorization checked before it occurs.
Reuse an established team playbook
A team playbook can package reviewed roles and instructions for another environment. Include what the receiving team needs to understand the method, then reconnect services and approve execution there. Keep existing credentials out of the package so reusable task knowledge does not carry one environment's access into another.
Inspect a portable team package
Before reusing a team playbook in another environment, examine the roles, instructions, references, and any included task material. Identify which services need fresh connections and which actions still require consent from the receiving user. Keep credentials and machine access out of the assumed handoff so the reusable package transfers a reviewed method rather than one environment's authority.
Prepare MausBot engines and connections
A working agent engine, computer or service access, and a reviewed team package are separate prerequisites. Check the route needed by the task and the receiving environment's permissions; a shared playbook does not supply credentials or unlimited model access.
Confirm a working agent engine
The desktop setup checks a supported local engine and its account. An open-source application does not include unlimited model access or remove provider limits.
Choose the computer and connections
Local computer use, cloud desktops, and connected applications require different permissions and services. Test the specific path needed by the task.
Check approvals and portable data
Review what the team file contains and which actions require consent. A shared playbook should not be treated as a transfer of credentials or computer access.
Related Software Projects
Compare the workflows, operating requirements, and source licenses of these projects.
MausBot FAQ
Are MausBot and OpenMausBot different projects?
The upstream README identifies MausBot as another name for OpenMausBot and records OpenGrokBot as a former name. The canonical source used here is milind-soni/OpenMausBot.
Is MausBot a language model?
No. It is an application and local harness around supported agent engines. Bot roles and model choices sit above those engines; inference still depends on the selected provider.
Does local-first mean all work stays offline?
No. Local conversations and workspace data coexist with requests to model providers, connected applications, or cloud computers when those features are used. Evaluate each enabled service separately.
What makes a channel different from a single bot thread?
A channel can hold shared instructions, a working folder, multiple bots, and responder rules. It defines the context in which those roles collaborate.
Can a shared team package carry live credentials?
The documented team-sharing flow excludes keys, passwords, chat history, model choices, and computers. Import behavior differs for organization-managed packages, so inspect the relevant sharing documentation.
Is the open-source app the same as free model and cloud access?
No. Apache-2.0 covers the application source. Agent accounts, API usage, connected services, cloud computers, and managed offerings have separate costs and terms.