Sub-Agents

User-built mini-agents that Faire calls as tools for focused, token-efficient task delegation.

Configured Sub-Agents

	Borrowed		Tools	Actions
get-uptime disabled `get-uptime`	Grian	read-only	run_command deliver	Edit
Report Scaire Github Updates disabled `scairegithubupdates`	Grian	read-only	run_command 26 MCP deliver	Edit
S.C.A.I.R.E Wiki Reader disabled `scairewikireader`	Grian	read-only	run_command 15 MCP deliver	Edit

Overview

Sub-Agents

0 active / 3 paused

Security Tier

3 RO

All read-only

Tools

3 cmd 2 MCP

Have delegated tools

What are Sub-Agents?

Sub-agents are focused, short-lived mini-agents that Faire calls as tools for token-efficient task delegation. Each one has its own instructions, tools, model, and budget. Faire delegates a task, the sub-agent runs it, and returns the result.

Inherits security at the borrowed agent's tier or below. Never above.
Composes its own context from prompt shards, free-text, and an immutable security block.
Calls deliver() to hand its result back to Faire.
Can use its own LLM provider and model, or inherit from the borrowed agent.

Sub-agents are optional. Faire works fine without them.

They shine for repetitive, specialized tasks: log analysis, config audits, structured data extraction. Build one when you notice Faire doing the same multi-step investigation over and over.