Agent Presets

Noeta ships four official agents aligned with Claude Code's roster. The agent is chosen per task in the POST /tasks body ({"goal": …, "agent": …}), not at process launch. Custom agents go through the flat Options.agents dict.

The quartet

Agent	Role	Tools	Capabilities
main	Default coding agent: full built-in tool surface + can spawn the three subagents + all capabilities.	Full built-in set (all fs + web + app + memory tools)	todo_write, ask_user_question, delegation, skill_invocation, memory, mcp
general-purpose	Self-contained coding worker: full read/write/edit/shell set, no delegation.	apply_patch, edit, glob, grep, read, shell_kill, shell_poll, shell_run, web_search, webfetch, write	skill_invocation, mcp
explore	Read-only scout: glob/grep/read + read-only shell, fans out to report facts, never edits.	glob, grep, read, shell_kill, shell_poll, shell_run, webfetch	skill_invocation
plan	Read-only architect: reads the code and returns a concrete ordered implementation plan, never writes.	glob, grep, read, shell_kill, shell_poll, shell_run, webfetch	ask_user_question

Capability flags

Flag	What it enables
todo_write	The todo_write control tool (state-patch based progress tracking).
ask_user_question	The model can yield for human input via ask_user_question control tool.
delegation	Can spawn subtasks via the three subagents (general-purpose, explore, plan).
skill_invocation	The skill control tool for model-driven skill selection.
memory	Cross-task memory: memory_write / memory_read tools + auto-recall at user-message seam. Only main enables this.
mcp	MCP tool inheritance: subtasks whose own spec also opens mcp inherit the parent's enabled MCP servers.

Sub-agent fan-out

main can spawn the three subagents in parallel; the result is the subagent's return value, recorded into the EventLog so the whole tree folds back into state. See ADR: Subtask fan-out and durable wake and ADR: Subtask parallel execution.

Using presets programmatically

from noeta import presets
from noeta.sdk import Client, query

# Build the main agent's Options
options = presets.main_options()

# Run an agent in-process
result = query(options, goal="Refactor module X to use Y")

Or compile all four agents as specs:

from noeta.presets import official_specs
specs = official_specs()
# → {"main": AgentSpec, "general-purpose": AgentSpec, "explore": AgentSpec, "plan": AgentSpec}

Custom agents

Define custom agents via the flat Options.agents dict:

from noeta.sdk import Options, AgentDefinition

options = Options(
    system_prompt="You are a docs writer.",
    agents={
        "reviewer": AgentDefinition(
            description="Reviews docs for accuracy and clarity.",
            prompt="...",
            tools=["read", "grep", "glob"],
        ),
    },
)

Source

• Presets: packages/noeta-runtime/noeta/presets/__init__.py
• Options / AgentDefinition: packages/noeta-sdk/noeta/client/options.py
• Tool catalog: packages/noeta-runtime/noeta/tools/
• See also: ADR: Tool and agent catalog, ADR: Library-SDK architecture