メニュー
Submit development specs, approve execution plans, monitor workers, retry failed tasks, check job status, and manage orchestration jobs. Use when asked to start a development job, check worker progress, approve a plan, retry something, or list orchestrator jobs.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Read and control I2C and SPI peripherals on Sipeed boards (LicheeRV Nano, MaixCAM, NanoKVM).
Generate synthetic training data, manage seeds, preview synthesis jobs, and work with the data factory. Use when asked to generate synthetic data, create training examples, augment a dataset, manage seeds, or check factory job status.
Browse, compare, evaluate, and manage trained models. View leaderboard, trends, lineage, and artifacts. Deploy models to Ollama. Use when asked about models, model comparison, which model is best, model evaluation, deploying a model, or downloading a model.
Check BashGym system status, GPU utilization, hardware info, and aggregate statistics. Use when asked about system health, what's running, GPU usage, memory, disk space, or general system stats.
Browse, promote, demote, classify, and generate training examples from Claude Code execution traces. Use when asked about traces, gold traces, trace counts, promoting or demoting traces, generating examples, exporting training data, or syncing traces.
Start, stop, pause, resume, and monitor model training runs. Check training progress, view metrics, and get GPU utilization during training. Use when asked to train a model, check training status, view loss curves, stop a run, or see training logs.
| name | orchestrator |
| description | Submit development specs, approve execution plans, monitor workers, retry failed tasks, check job status, and manage orchestration jobs. Use when asked to start a development job, check worker progress, approve a plan, retry something, or list orchestrator jobs. |
Submit development specs for multi-agent decomposition and parallel execution.
Decompose a development task into a parallel execution plan:
scripts/api.sh POST /orchestrate/submit '{
"title": "Add user authentication",
"description": "Implement JWT-based auth with login/register endpoints",
"constraints": ["Use existing User model", "Add pytest tests"],
"acceptance_criteria": ["All tests pass", "Endpoints return proper status codes"],
"max_workers": 3,
"budget_usd": 5.0,
"provider": "anthropic"
}'
Returns: job_id and decomposed task plan. Always summarize the plan for the user before approving.
After reviewing the plan, approve to begin execution:
scripts/api.sh POST /orchestrate/{job_id}/approve '{}'
Monitor worker progress and task states:
scripts/api.sh GET /orchestrate/{job_id}/status
Returns: task states (pending/running/completed/failed), worker assignments, budget usage.
Retry a specific failed task within a job:
scripts/api.sh POST /orchestrate/{job_id}/task/{task_id}/retry '{}'
scripts/api.sh GET /orchestrate/jobs
Destructive — confirm with user first.
scripts/api.sh DELETE /orchestrate/{job_id}
Check which LLM providers are configured for planning:
scripts/api.sh GET /orchestrate/providers
User: "Build me an auth system with JWT" → Ask for any constraints. Compose a spec. Call POST /orchestrate/submit. Summarize the decomposed plan. Ask user to approve before calling /approve.
User: "What's the status of my last job?" → Call GET /orchestrate/jobs to find the latest. Call GET /orchestrate/{job_id}/status. Report task states and budget usage.
User: "Task 3 failed, retry it" → Call POST /orchestrate/{job_id}/task/{task_id}/retry. Report result.