Skip to content

LLMKube

v0.9.7 Feature

This release adds 3 notable features for engineering teams evaluating rollout.

✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

ai apple-silicon autoscaling edge-computing gguf gpu
+12 more
self-hosted inference kubernetes llama-cpp llm local-llm metal mlx multi-gpu nvidia tgi vllm

Summary

AI summary

Updates 0.9.7, Bug Fixes, and 2026-07-16 across a mixed release.

Changes in this release

Feature Medium

Adds support for pinned multi-file Hugging Face models.

Adds support for pinned multi-file Hugging Face models.

Source: llm_adapter@2026-07-16

Confidence: high

Feature Medium

Exposes oMLX paged SSD cache configuration in metal-agent.

Exposes oMLX paged SSD cache configuration in metal-agent.

Source: llm_adapter@2026-07-16

Confidence: high

Feature Medium

Adds router in-memory rolling-window budget accounting engine.

Adds router in-memory rolling-window budget accounting engine.

Source: llm_adapter@2026-07-16

Confidence: high

Feature Medium

Guides coders to verify volatile external facts via MCP research tools in foreman.

Guides coders to verify volatile external facts via MCP research tools in foreman.

Source: llm_adapter@2026-07-16

Confidence: high

Feature Medium

Adds opt‑in coder escalation on capability‑wall failures in foreman.

Adds opt‑in coder escalation on capability‑wall failures in foreman.

Source: llm_adapter@2026-07-16

Confidence: high

Feature Medium

Adds cli envtest‑verify for the slicer's integrated branch.

Adds cli envtest‑verify for the slicer's integrated branch.

Source: llm_adapter@2026-07-16

Confidence: high

Feature Medium

Adds foreman-finalize.sh post‑GO branch finalizer script.

Adds foreman-finalize.sh post‑GO branch finalizer script.

Source: llm_adapter@2026-07-16

Confidence: high

Bugfix Medium

Fixes foreman aborting orphaned agent loop when its AgenticTask is deleted.

Fixes foreman aborting orphaned agent loop when its AgenticTask is deleted.

Source: llm_adapter@2026-07-16

Confidence: high

Bugfix Medium

Fixes foreman-finalize script to handle Foreman CRDs, clean‑tree, and title override.

Fixes foreman-finalize script to handle Foreman CRDs, clean‑tree, and title override.

Source: llm_adapter@2026-07-16

Confidence: high

Full changelog

0.9.7 (2026-07-16)

Features

  • cli: envtest-verify the slicer's integrated branch (Fixes #1137) (#1141) (a7622c2)
  • foreman: guide coders to verify volatile external facts via MCP research tools (#1134) (390c945)
  • foreman: opt-in coder escalation on capability-wall failures (#1108) (#1128) (68e2211)
  • metal-agent: expose oMLX paged SSD cache configuration (#278) (#1129) (095e4a2)
  • router: in-memory rolling-window budget accounting engine (#434) (#1130) (c6ba850)
  • scripts: add foreman-finalize.sh post-GO branch finalizer (#1126) (8d17e25)
  • Support pinned multi-file Hugging Face models (#1110) (#1138) (e2d3441)

Bug Fixes

  • foreman: abort orphaned agent loop when its AgenticTask is deleted (Fixes #1136) (#1140) (471a00c)
  • scripts: foreman-finalize handles Foreman CRDs, clean-tree, and title override (#1131) (db3eeaf)

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Track LLMKube

Get notified when new releases ship.

Sign up free

About LLMKube

Kubernetes operator for llama.cpp-native LLM inference with GPU scheduling, Apple Silicon Metal support, and OpenAI-compatible API.

All releases →

Related context

Earlier breaking changes

  • v0.8.1 foreman: requestTimeoutSeconds now sets loop-wide budget, default changes from 600 to 3600.

Beta — feedback welcome: [email protected]