SeaSnoke

Docs

Core Concepts

Agent System

How SeaSnoke's coding agents work, from initialization to output validation.

Agent Lifecycle

Each agent follows a well-defined lifecycle: Initialization, Planning, Execution, Validation, and Reporting. Agents are stateless by default but can maintain context within a single task run.

Agent Types

  • Architect

    Designs system structure, APIs, and data models before implementation.
  • Implementer

    Writes the actual code, tests, and documentation based on the architecture.
  • Reviewer

    Analyzes code for bugs, security issues, and style violations.
  • Test Engineer

    Generates comprehensive test suites, including edge cases and property tests.