Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Skills for spawning external processes - AI coding agents and generic CLI commands in new terminal windows. Parent skill category for agent and terminal spawning.
.claude/skills/majiayu000-spawn/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -1% | 0% |
This directory contains skills for spawning external processes in new terminal windows.
Spawn skills enable Claude Code to launch external processes:
Both use the fork_terminal utility to create isolated terminal sessions.
| Skill | Description | Use Case | |-------|-------------|----------| | agent | Spawn AI coding agents | Multi-provider orchestration | | terminal | Spawn generic CLI commands | Non-AI command execution |
Both skills use the fork_terminal Python utility:
python# Located at: ./agent/fork_terminal.py from fork_terminal import fork_terminal # Basic usage result = fork_terminal("command", capture=True) # With logging result = fork_terminal("command", log_to_file=True, log_agent_output=True)
spawn/
├── SKILL.md # This file
├── agent/ # AI agent spawning
│ ├── SKILL.md
│ ├── cookbook/ # Per-agent cookbooks
│ │ ├── claude-code.md
│ │ ├── codex-cli.md
│ │ ├── gemini-cli.md
│ │ ├── cursor-cli.md
│ │ ├── opencode-cli.md
│ │ └── copilot-cli.md
│ └── prompts/ # Reusable prompt templates
└── terminal/ # Generic CLI spawning
├── SKILL.md
└── cookbook/
└── cli-command.md.claude/ai-dev-kit/dev-tools/orchestration/providers/ - Shell scripts for each provider/ai-dev-kit:delegate - Command for manual delegation| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 15,839 | 10,083 | -36% | 1 | 1 | 0% | 1,520 | 1,473 | -3% | 0 | 0 | — |
case-02 | fail→pass | 9,903 | 9,778 | -1% | 1 | 1 | 0% | 1,611 | 1,385 | -14% | 0 | 0 | — |
case-03 | fail→pass | 29,737 | 8,644 | -71% | 1 | 1 | 0% | 1,608 | 1,266 | -21% | 0 | 0 | — |
case-04 | fail→pass | 14,860 | 7,964 | -46% | 1 | 1 | 0% | 2,289 | 1,098 | -52% | 0 | 0 | — |
case-05 | fail→pass | 16,449 | 10,498 | -36% | 1 | 1 | 0% | 1,656 | 1,636 | -1% | 0 | 0 | — |
case-06 | fail→pass | 23,792 | 2,841 | -88% | 1 | 1 | 0% | 4,164 | 1,164 | -72% | 0 | 0 | — |
case-07 | fail→pass | 7,271 | 7,916 | +9% | 1 | 1 | 0% | 1,166 | 1,123 | -4% | 0 | 0 | — |
case-08 | pass→pass | 20,133 | 10,925 | -46% | 1 | 1 | 0% | 2,460 | 1,719 | -30% | 0 | 0 | — |
case-09 | pass→pass | 16,549 | 4,217 | -75% | 1 | 1 | 0% | 3,102 | 1,348 | -57% | 0 | 0 | — |
case-10 | pass→pass | 19,364 | 12,456 | -36% | 1 | 1 | 0% | 2,498 | 1,925 | -23% | 0 | 0 | — |
case-11 | fail→pass | 17,579 | 2,810 | -84% | 1 | 1 | 0% | 2,094 | 1,077 | -49% | 0 | 0 | — |
case-12 | fail→pass | 13,327 | 10,602 | -20% | 1 | 1 | 0% | 1,316 | 1,126 | -14% | 0 | 0 | — |
case-13 | fail→pass | 9,587 | 9,712 | +1% | 1 | 1 | 0% | 762 | 1,471 | +93% | 0 | 0 | — |
case-14 | fail→pass | 13,376 | 8,536 | -36% | 1 | 1 | 0% | 2,123 | 1,191 | -44% | 0 | 0 | — |
case-15 | fail→pass | 11,469 | 3,088 | -73% | 1 | 1 | 0% | 1,860 | 1,207 | -35% | 0 | 0 | — |
case-16 | fail→pass | 12,072 | 4,023 | -67% | 1 | 1 | 0% | 2,037 | 1,397 | -31% | 0 | 0 | — |
case-17 | fail→pass | 11,125 | 10,488 | -6% | 1 | 1 | 0% | 1,025 | 1,580 | +54% | 0 | 0 | — |
case-18 | fail→pass | 18,503 | 11,212 | -39% | 1 | 1 | 0% | 2,063 | 1,702 | -17% | 0 | 0 | — |
case-19 | pass→pass | 8,794 | 3,755 | -57% | 1 | 1 | 0% | 1,263 | 1,157 | -8% | 0 | 0 | — |
case-20 | fail→pass | 13,876 | 8,128 | -41% | 1 | 1 | 0% | 1,380 | 1,170 | -15% | 0 | 0 | — |
case-21 | fail→pass | 21,368 | 8,109 | -62% | 1 | 1 | 0% | 2,551 | 1,181 | -54% | 0 | 0 | — |
case-22 | fail→pass | 16,471 | 6,632 | -60% | 1 | 1 | 0% | 2,063 | 1,759 | -15% | 0 | 0 | — |
case-23 | fail→pass | 7,823 | 9,114 | +17% | 1 | 1 | 0% | 1,230 | 1,396 | +13% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +83 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.