[andypalmi]

🤖 New release

• amicu-core: 0.1.4 -> 0.1.5
• amicu-registry: 0.1.4 -> 0.1.5
• amicu-runtime: 0.1.4 -> 0.1.5
• amicu-llamacpp: 0.1.4 -> 0.1.5
• amicu-providers: 0.1.4 -> 0.1.5
• amicu-gcp: 0.1.4 -> 0.1.5
• amicu-hcl: 0.1.4 -> 0.1.5
• amicu-vllm: 0.1.4 -> 0.1.5
• amicu-engine: 0.1.4 -> 0.1.5
• amicu-recommend: 0.1.4 -> 0.1.5
• amicu-state: 0.1.4 -> 0.1.5
• amicu-store: 0.1.4 -> 0.1.5
• amicu-modal: 0.1.4 -> 0.1.5
• amicu-runpod: 0.1.4 -> 0.1.5
• amicu-cli: 0.1.4 -> 0.1.5
• amicu-auth: 0.1.4 -> 0.1.5
• amicu-telemetry: 0.1.4 -> 0.1.5
• amicu-api: 0.1.4 -> 0.1.5
• amicu-server: 0.1.4 -> 0.1.5
• amicu-worker: 0.1.4 -> 0.1.5

Changelog

amicu-cli

0.1.5 - 2026-10-28

Added

• (gcp) network and subnetwork options for Compute Engine (#369)
• (cli) deploy to Cloud Run GPUs with amicu apply (#361)
• (cli) deploy to Compute Engine GPU VMs with amicu apply (#354)
• (runpod) network-volume model cache, flashboot and live test (#333)
• (runpod) Pods in refresh, orphans and runpod commands (#328)
• (runpod) always-on Pods mode (#324)
• (runpod) status, logs and metrics (#320)
• (runpod) reconcile RunPod in refresh and orphans (#319)
• (runpod) pod client and PodSupervisor (#315)
• (runpod) wire RunPod into apply with write-ahead state (#316)
• (hardware) shared GPU catalog with exact variant names (#312)
• (runpod) supervisor and gateway auth (#309)
• (runpod) crate skeleton and v2 REST client (#307)
• (providers) Cloud Run GPU service lifecycle for GCP (#352)
• (providers) Secret Manager secrets and pinned TLS for GCP VMs (#353)
• (providers) Compute Engine GPU VM lifecycle for GCP (#351)
• (providers) GCP GPU catalog by zone with a live drift check (#350)
• (providers) add amicu-gcp core crate (auth, REST client, operations, owner labels) (#348)

Fixed

• (gcp) install Docker on the VM at boot and fail apply fast on a failed startup script (#363)
• (gcp) send a body with every POST and build GCP clients inside the runtime (#362)
• (providers) GCP preflight, secret teardown and idle shutdown (#360)
• (cli) run the telemetry agent for Modal and RunPod deployments (#349)
• (gcp) list GPU zones one by one in the catalog drift check (#370)
• (runpod) raise the serverless startup timeout for long model loads (#337)
• (runpod) require the key on every vLLM pod route (#332)
• (runpod) plan shows the remote launch and placement (#331)
• (runpod) llama.cpp routing and live-test follow-ups (#330)
• (runpod) region placement, llama.cpp health and CUDA floor, readiness progress (#325)
• (gcp) sync the GPU catalog with the live Compute Engine API (#366)

Other

• (deps) bump russh from 0.63.2 to 0.64.1 (#346)

---

This PR was generated with release-plz.