# Agent Spawner

> Run one isolated delegated agent from the AgentConfig model, MCP, TLS, and skills catalogs. Use when a bounded task needs a temporary specialist without mutating the parent agent; prefer GraphOS delegation for graph-governed work.

- Skill: `knuckles-team/agent-spawner` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add knuckles-team/agent-spawner`
- Raw SKILL.md: https://api.skillmd.com/api/skills/knuckles-team/agent-spawner/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- License: MIT
- Author: Knuckles-Team (https://skillmd.com/u/knuckles-team)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/knuckles-team/agent-spawner

---


# Agent Spawner

Run one temporary isolated agent using deployment-neutral `AgentConfig`:

```bash
python scripts/spawn_agent.py --prompt "Inspect the configured tool catalog"
```

`CHAT_MODELS`, `MCP_CONFIG`, `MCP_URL`, the custom-skills root, TLS policy, and
runtime secret resolution come from AgentConfig/XDG configuration. Provider and
model identifiers may be selected explicitly when they already exist in the
deployment's model catalog:

```bash
python scripts/spawn_agent.py \
  --provider openai \
  --model-id configured-model \
  --prompt "Perform the bounded delegated task"
```

The compatibility flags that accepted API keys, arbitrary dotenv/config paths,
base URLs, personal agent names, or disabled TLS verification are retired.
Never put credentials in arguments. The child uses strict tool guards and MCP
isolation, emits no prompt/name/path/endpoint details, and relies on the shared
transport profile. Use GraphOS delegation when policy, capability provenance,
or trace-linked skill execution is required.