@deepseek-ai/dsh-agent-teams

Configuration

Key	Default	Meaning
provider	spawn	The ctx.subagents provider role subagents run on.
toolName	team	The model-facing tool name.
maxConcurrentRoles	4	Maximum concurrently running role subagents.
maxRoleOutputChars	20000	Per-role output ceiling in characters.
maxRoles	8	Maximum roles per run.
maxDepth	3	Absolute delegation-depth cap for every child.
enableRunInBackground	true	Exposes run_in_background on the tool.
maxResultChars	50000	Rendered-result ceiling in characters.
templates	[]	Extra team templates merged over the builtins (ids must be unique).

team usage guidance section

What the model sees

A fixed prompt section is registered while the tool is loaded, telling the model to use team only when the user explicitly asks for a multi-agent team or a task that genuinely splits into concurrent specialized roles, and to prefer plain subagent calls for one or two delegations.

Verbatim text for this section

Use the team tool ONLY when the user explicitly asks for a multi-agent team or a task that genuinely splits into several concurrent specialized roles. For one or two delegations, prefer plain subagent calls.

Token effect

Conditional: the section text is appended to the system prompt while the tool is registered.

KV Cache effect

Replacing: the section is a new prompt contribution that changes the assembled request prefix when the plugin loads or unloads; while loaded it is stable and does not itself invalidate reuse.