AI & agents · Automation
team-spawn
/team-spawn
Use when creating a confirmed Codex native subagent team for bounded parallel work.
For whom
- Author
- hoya324 ↗
- Repository
- oh-my-harness ↗
- Licence
- MIT
- Last repository change
- 28.07.2026
Assessment and security
100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
Deprecation
No signals detected. Repository changed on 28.07.2026, without references to retired models or APIs.
Licence
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Use when creating a confirmed Codex native subagent team for bounded parallel work.
Original author description
Use when creating a confirmed Codex native subagent team for bounded parallel work.
How it works
- Resolve the team
- Confirmation gate
- Spawn and persist
Original SKILL.md headings · 85 lines, 467 words
SKILL.md preview
---
name: team-spawn
description: Use when creating a confirmed Codex native subagent team for bounded parallel work.
---
# Spawn a Codex native team
Read `../../references/runtime-map.md`. Never spawn agents during installation, setup, discovery,
or merely because this skill loaded.
## Resolve the team
Parse a template name (`fullstack`, `review`, `research`), positive teammate count, or custom
composition plus a concrete task. Read `.claude/.omh/harness.config.json`:
- stop when `features.nativeTeam` is false;
- enforce `nativeTeam.maxTeammates`;
- use `nativeTeam.defaultTeamName` and configured templates;
- treat `modelRouting` and agent profiles as preferences, not availability guarantees.
Require the team name to match `^[a-zA-Z0-9_-]+$`; ask for another name instead of silently
changing user input. Treat any existing `.claude/.omh/teams.json` as protected, including
`starting`, active, partial, malformed, failed, stopped-but-unclean, or otherwise unresolved state.
Call `list_agents` and reconcile exact persisted canonical tasks and opaque agent ids when parsing
is possible. Preserve malformed raw state for recovery.
Present cleanup and resume choices and require an explicit decision before any replacement:
- `resume`: preserve the file and recovery ids, then continue only the reconciled existing team;
- `cleanup`: use `/team-stop` behavior and wait for observed terminal agents before changing state;
- `cancel`: change nothing.
Never overwrite recovery ids. Replacement requires explicit confirmation after cleanup is observed
complete; a malformed or partial file is not permission to initialize a new team.
Ask necessary questions directly in chat when “review,” “fix,” or another goal lacks files,
acceptance criteria, or compatibility boundaries. Decompose the task into concrete, bounded
assignments with explicit deliverables, paths, verification, and dependencies. Spawn only
assignments that can run now; retain blocked work in state for later dispatch. Avoid multiple
agents editing the same files unless the user accepts the collision plan.
## Confirmation gate
Show team name, task, template, every agent name/role, exact assignment, dependency, proposed model
when known, and state path. Explain that each approval creates a live Codex subagent. Obtain
explicit confirmation before any state write or `spawn_agent` call. Cancel without side effects on
any answer other than clear approval.
## Spawn and persist
After confirmation, create `.claude/.omh/teams.json` atomically with `status: "starting"` and the
confirmed task records. Call `spawn_agent` once per ready assignment using a stable unique task
name. Do not invent or normalize returned identifiers. After each success, atomically persist the
returned canonical task name and agent id so partial failure is recoverable:
```json
{
"teamName": "omh-team",
"createdAt": "ISO timestamp",
"description": "confirmed task",
"template": "review",
"status": "active",
"teammates": [
{
"name": "reviewer",
"role": "Code reviewer",
"agentType": "architect",
"canonicalTask": "/root/reviewer",
"agentId": "opaque returned id",
"assignment": "bounded assignment",
"status": "working"
}
],
"tasks": [
{
"name": "reviewer",
"subject": "Review payment refactor",
"status": "in_progress",
"blockedBy": []
}
]
}
```
If a spawn fails, stop dispatching, retain confirmed state and successful ids, and report the
partial team. Use `send_message` only with a persisted canonical task or agent target when
coordination is needed. Never claim a requested profile proves a particular model. Report
`/team-status` and `/team-stop` as next actions.
Versions
The database has no other skills named team-spawn.
Database snapshot: 13.11.2026 · package version 698ac0dc50c2 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100agent-team
This skill should be used when the user asks to "create a team", "spawn agents", "parallel execution", "delegate tasks to agents", "use a swarm", "team workflow", "split work across agents", "run tasks in parallel", or mentions wanting multiple agents to collaborate on a complex task. - 100spawn-ours-agent
Spawn and configure a new ours-fleet agent from Codex, with a permanent or temporary lifetime and an ACP session. Use when the user asks to spawn an agent, create a fleet role, start a background agent, delegate work to another Codex session, choose its model or permissions, or… - 100codex-subagent
Launch OpenAI Codex CLI as a subagent (ChatGPT subscription auth, no API key). Use when delegating a self-contained coding task to Codex from another agent — parallel implementation work, a second opinion, or an independent verification pass. - 100agent-teams
Create agent teams for parallel collaboration. Use when user mentions team, parallel, multi-agent, coordinate, collaborate, review team, multiple perspectives, or says "agent-teams". - 100team-chat
Work as a team with the other coding agent (Claude <-> Codex) through Team Chat. Use when the user says "ask Codex", "ask Claude", "have the other agent do X", "get ChatGPT/Codex to make images", "tell me when it's done", "check your team inbox", "team chat", or when a task… - 86team-stop
Use when interrupting a Codex native subagent team or cleaning its shared OMH state.