Overview
In Ottili Coder you pick a mode* before starting a mission. The mode is a clear work instruction: it defines what Coder is allowed to do – from a pure question to a complete feature implementation. Coder automatically appends the matching instruction to your mission so the model performs exactly the job you intended.
The five core modes are Ask, Plan, Build, Fix and Review*. They are complemented by Stabilize, Deploy and Full Run* for recurring and broader tasks. All modes share the same mission, the same task queue, the same logs and the same audit trail – they differ only in what may be changed and what outcome Coder targets.
The modes are available in the Ottili Coder Desktop app* (status: available / AVAILABLE). The actual execution runs locally, in the cloud or hybrid – cloud runs are currently in beta (BETA).
The five core modes
Ask
Ask is the pure question and analysis mode. Coder changes no files; it answers your question, explains code, finds risks or summarizes the current state.
Use Ask when you:
- want to understand the code or a behavior before changing anything,
- need an assessment without files being modified,
- want to close a knowledge gap (e.g. "How is this data flow structured?").
Plan
Plan is read-only planning. Coder inspects the code, proposes a structured, reviewable plan and writes plan files – but it does not modify your code.
Use Plan when:
- you want to understand the approach before anything changes,
- the change is large or risky,
- you have an implementation plan to review with your team.
You can promote the approved plan to a build run on the same mission. The queue, logs and audit trail stay coherent across the switch.
Build
Build is the default execution mode. Coder plans the work, runs agents, applies changes to your repository and can open a pull request. Validation and approval gates still apply, but the goal is a finished, reviewable result.
Use Build when you want Coder to do the work and hand back a branch or PR.
Fix
Fix is the bug-fix mode. Coder finds the root cause of a problem and fixes it. It is built for root-cause analysis and targeted correction, not for new features.
Use Fix when:
- a concrete error or bug exists,
- a test is failing and the cause is unclear,
- you want "why is X broken?" answered before larger work begins.
Review
Review is the inspection mode. Coder analyzes the code thoroughly and reports findings with severity – without making changes. This is the basis for human reviews and for approval gates.
Use Review when:
- you want a complete code and stability review of the repository,
- an open PR or change should be evaluated independently,
- security and quality risks must surface before they are merged.
Additional modes
Beyond the five core modes, Coder offers three complementary modes for recurring and broader tasks.
Stabilize
Stabilize runs tests, fixes failures and cleans up. It is the right mode to bring an unstable codebase back to green step by step.
Deploy
Deploy prepares a cloud deployment. Direct Ottili Cloud deployment is currently planned (PLANNED); the mode produces the deployment preparation, and release stays approval-first.
Full Run
Full Run is the fully autonomous pass: plan, build, test and verify in a single run. It combines the core modes into one continuous, validated result.
Modes and execution
The chosen mode only determines what Coder is allowed to do – not where the work runs. You launch the same mode locally, in Ottili Cloud* (beta) or hybrid. Regardless of mode, validation loops and – when configured – security checks apply; risky steps surface as approval gates instead of running silently.
Status and availability
| Mode | Modifies files | Status |
|---|---|---|
| Ask | No (analysis only) | Available (AVAILABLE) |
| Plan | No (planning only) | Available (AVAILABLE) |
| Build | Yes | Available (AVAILABLE) |
| Fix | Yes (targeted correction) | Available (AVAILABLE) |
| Review | No (findings only) | Available (AVAILABLE) |
| Stabilize | Yes | Available (AVAILABLE) |
| Deploy | Preparation (deployment planned) | Available (AVAILABLE), deployment planned (PLANNED) |
| Full Run | Yes | Available (AVAILABLE) |
The modes themselves are available (AVAILABLE) – the cloud execution* of any mode is beta (BETA), local execution is available. Direct cloud deployment capability is planned (PLANNED). How Ottili ONE labels maturity publicly is described in [Understand feature status labels](/docs/understand-feature-status-labels).
Related articles
- [Ottili Coder overview](/docs/ottili-coder)
- [Missions](/docs/coder-missions)
- [Task queues](/docs/coder-task-queues)
- [Agents](/docs/coder-agents)
- [Validation](/docs/coder-validation)
- [Local, cloud and hybrid runs](/docs/coder-local-cloud-hybrid-runs)
- [Security](/docs/coder-security)
- [Understand feature status labels](/docs/understand-feature-status-labels)
Was this article helpful?
