Claude Skills by jvogan

Use for Symphony + Linear cloud-compute work: launch manifests, provider setup guidance, local dry-runs, guarded remote runs where supported, artifact proof, cleanup, self-learning ledger, and symphony-outcome closeout.

Guardrails for AI agents that launch paid workloads on RunPod from Linear issues. Validates the launch manifest, runs a local dry-run, gates remote pod creation on explicit authorization, requires SHA-256 artifact proof for success, and forces cleanup. Use when a Symphony, Codex, or Claude Code worker is about to create a RunPod pod, when a remote run is stuck, or when closing out a RunPod job with symphony-outcome.