foxrun-ai/foxrun / skillwright/skills/agent-new
agent-new
VerifiedCreate or update an agent worker from any runtime: write its identity file, render its Claude and Codex capes, and list the team. Use when the person or a lead says new agent, new worker, new specialist, add a role, update an agent, or rebuild a cape.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 2.7 KB |
| bin/agent-newexecutable | script | 380 B |
| replays/list-help.json | config | 297 B |
| replays/read-seat-boot.json | config | 342 B |
| tests/__init__.py | script | 0 B |
| tests/test_wrapper.py | script | 1.1 KB |
SKILL.md
View on GitHub---
name: agent-new
description: "Create or update an agent worker from any runtime: write its identity file, render its Claude and Codex capes, and list the team. Use when the person or a lead says new agent, new worker, new specialist, add a role, update an agent, or rebuild a cape."
kind: tool
---
# agent-new
One identity file per agent. The capes are thin loaders rendered from it, so you change the identity, never the cape.
The wrapper `bin/agent-new` runs the installed program `~/.foxrun/agents/tools/agent_new.py`. Run `bin/agent-new <command> -h` for each command's flags.
## Where things go
- A workspace agent: `~/foxrun/<workspace>/team/<id>.md`. Its capes sit in `<workspace>/.claude/agents/<id>.claude.md` and `<workspace>/.codex/agents/<id>.toml`; both folders link to `<workspace>/team`.
- A system agent: `--client system`. Its capes sit in `~/.foxrun/agents/generated/claude/<id>.md` and `~/.foxrun/agents/generated/codex/<id>.toml`.
- Render receipts: `<workspace>/history/outputs/agent-new/<id>/`, or `~/.foxrun/agents/generated/<id>/` for a system agent.
- An id has one owner across both homes. A duplicate id is refused.
## Steps
1. Check the id is free: `bin/agent-new list`. For an existing seat, run `python3 ~/.foxrun/foxrun.py start --seat <id>` and read the file on its `File:` line. `base.paths.resolve_seat` selects that file.
Done when: the id is free (create), or you have read its current identity file (update).
2. Create: run with `--dry-run` first, then without it.
```
bin/agent-new create <id> --role "<role>" --summary "<one line>" --tier work --skills a,b --tools qmd --client <workspace> --runtimes claude,codex --hermes-mode none --dry-run
```
Update: edit the identity file, bump its `version`, and go to step 3.
Done when: the identity file exists with the right role, tier, skills, and version.
3. Render the capes: `bin/agent-new render <id> --dry-run`, then `bin/agent-new render <id>`.
Done when: each cape names the exact identity path and carries the `managed-by: agent-new` marker.
4. Report the identity path, each cape path, and the receipt folder.
Done when: `bin/agent-new list` shows the id with its client.
## Gotchas
- An identity is not a Hermes profile. Only `--hermes-mode none` works today. Child and profile modes are held.
- `probe`, `diff`, and `apply` stop today. They need an approved agent-folder plan and live loader proof. No stopped command falls back to an old home.
- Skills are ids on the shared shelf. Report a missing skill as a gap. Never invent one.
- New skill names follow `~/.foxrun/skillwright/standards/skill-law.md`.
- A hand-written cape has no marker. `render` never overwrites it.
- Tiers resolve to models in `~/.foxrun/config/models.yaml`.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |