Typed Decisions Go Local, Claude Code Gets Cheaper, MCP Shell Access Bites Back

Today's findings

1. #1 Disable Claude Code's Prompt Suggestions to save ~10% of usage limitstip

   The greyed-out inline "Prompt Suggestions" text in Claude Code triggers a full context cache read on every pause, quietly burning quota.

   Claude Code · cost optimization

   Greyed-out autocomplete quietly burns quota

   ~10%

   claimed saving from one settings toggle

   free fix

   Every pause

   fires a full context cache read

   Inline grey text

   the "Prompt Suggestions" hints

   $0

   cost to switch it off in settings

   Figure is the source's own estimate, not a measured benchmark.

   Why it matters: With Claude Code usage limits a constant pain point, a free settings toggle that recovers roughly a tenth of your spend/limit is a no-cost win.

   How to apply: Turn off Prompt Suggestions in Claude Code settings if you don't rely on the inline "commit and push"-style autocomplete text.

   claude-codecost-optimization

   Read more: PSA - Claude Code: Turn off Prompt Suggestions, save ~10% of your limits/spend

2. #2 Sandbox shell/file MCP tools before pointing local models at themtip

   Local models like Granite 4.2 8B will happily write files to absolute root paths (e.g. "/newPythonCode.py") when given unsandboxed shell/file MCP access.

   MCP · local coding agents

   Let the local model write files — only inside a box it cannot escape

   bounded capability

   Shell + file-write MCP tools

   scope Disposable container

   scope Chroot to project

   limit No leading slash

   monitor Inspect write paths

   Granite 4.2 8B happily wrote to "/newPythonCode.py" when the MCP server was unsandboxed.

   Why it matters: As more teams wire local models into coding agents via MCP shell servers, an absolute-path write outside a sandbox can silently overwrite files on the host rather than the intended project directory.

   How to apply: Run shell/file-write MCP servers inside a disposable container or chroot, and treat any leading-slash path in a model's file-write tool call as a red flag worth catching before execution.

   mcpagentslocal-llm

   Read more: an (unsafe) command shell mcp server

3. #3 Community loader fixes unlock ternary ~1.7-bit Bonsai-2-27B in LM Studio and MLXtool

   Prism ML's sub-2-bit ternary quant of Qwen3.8-27B (5.9–8.4GB) failed to load in stock LM Studio and oMLX until community patches added support for the custom ggml/MLX types.

   Ternary quant · Bonsai-2-27B

   The quant wasn't broken — the loaders were

   Stock runners

   • "invalid ggml type"
   • "model type not supported"
   • Custom ternary type unknown
   • Looks like a bad download

   → 27B at ~1.7 bits/weight → 5.9–8.4 GB on disk

   Community patches

   • lmstudio-prism-bonsai loader
   • oMLX fork adds MLX type
   • Ternary weights recognized
   • Loads on modest hardware

   Prism ML's sub-2-bit ternary Qwen3.8-27B: check for a loader patch before blaming the model.

   Why it matters: Ternary quantization squeezes a 27B model into a size that runs comfortably on modest hardware, but mainstream runners lag behind novel quant formats and fail with unhelpful errors.

   How to apply: If you hit "invalid ggml type" or "model type not supported" loading an exotic new quant, check for a community loader patch (lmstudio-prism-bonsai, oMLX fork) before assuming the quantized model itself is broken.

   quantizationggufmlxlocal-llm

   Read more: Ternary Bonsai 2 (27B) fails to load in LM Studio and oMLX. I made fixes for both (GGUF PQ2_0/PTQ1_0 + MLX 2-bit)

4. #4 A fully local 30B model opened a real GitHub PR via llamafile, no cloud involvedtechnique

   Muse Glimmer 30B packaged as a single llamafile, driven by the Hermes coding agent through the local server mode, diagnosed and opened an actual pull request against a live open-source bug with zero API keys.

   Local agentic coding

   From open weights to a merged-ready GitHub PR without leaving the machine

   1

   30B weights

   Muse Glimmer, open

   2

   One llamafile

   single-file package

   3

   Local server

   zero API keys

   The localhost endpoint stands in where a hosted API normally sits

   4

   Hermes agent

   diagnoses the bug

   5

   Real PR

   live OSS repo

   No cloud call, no vendor key — the whole agentic loop runs on local hardware.

   Why it matters: It's concrete evidence that local, open-weight models plus a llamafile + agent harness can now handle genuine agentic coding tasks end-to-end, without cloud API cost or a vendor dependency.

   How to apply: For cost-sensitive or air-gapped agentic coding experiments, package a capable local model as a llamafile and point a coding agent at its local server mode instead of defaulting to a hosted API.

   local-llmagentscoding

   Read more: We had a local 30B model open a real pull request, no API key, no cloud

5. #5 hearim: open-source gateway turns any local LLM into a typed-decision APItool

   hearim wraps your existing llama.cpp, vLLM, SGLang, or Ollama endpoint so it returns calibrated typed decisions with probabilities instead of free text, no fine-tuning or new model required.

   OPEN-SOURCE GATEWAY

   Two ways to get a decision out of a local LLM

   Free text, then parse

   • Model writes a sentence
   • You regex or JSON-parse it
   • Confidence uncalibrated
   • Extra tokens, extra latency

   vs

   Typed decision endpoint

   • Answer set defined up front
   • Probability per option
   • Calibrated, no parsing
   • No fine-tune, same server

   hearim wraps the endpoint you already run — the answer set becomes the API, not the prose.

   Fronts llama.cpp, vLLM, SGLang or Ollama for routing, relevance and moderation gates.

   Why it matters: Classification-style steps in an agent pipeline (routing, relevance scoring, moderation gates) are often served by full text generation you then parse — slower, costlier, and less calibrated than getting a probability distribution directly from a model you already run.

   How to apply: Point hearim at your existing local/hosted model server, define the allowed answer set, and swap out open-ended classification calls in your pipeline for its decision endpoint instead of building a proprietary decision-API integration.

   local-llmagentsclassification

   Read more: hearim(헤아림): Maybe you don’t need a special model for Jev — ordinary local LLMs already have the capability

6. #6 Speculative decoding drafting strategy for Qwen3.8 needs per-model tuningtip

   llama.cpp users report mixed results combining MTP/DFlash draft speculative decoding with n-gram map-k drafting on Qwen3.8 Flash-Next and 27B.

   Why it matters: Speculative decoding is one of the few free local-inference speedups available, but the optimal draft strategy is model- and workload-specific and isn't well documented anywhere.

   How to apply: If serving Qwen3.8 models locally, start with `spec-type: draft-mtp,ngram-map-k` for Flash-Next or `draft-dflash,ngram-map-k` for 27B with `spec-draft-n-max: 4`, then benchmark against your own workload rather than trusting default configs.

   speculative-decodingllama.cpplocal-llm

   Read more: Which speculative decoding are you using and why? · Which speculative decoding are you using and why?

7. #7 Qwen-Image-2.1 tops open-weight image editing benchmarks but drops Apache-2.0tool

   Qwen-Image-2.1 (7B) scores highest among open-weight models on Qwen's own image-editing benchmark, with community GGUF/uncensored variants and simple ComfyUI workflows already live, but it ships under a more restrictive license than its predecessor.

   Why it matters: It's immediately usable for local image-editing pipelines, but the license change matters if you planned to embed it commercially the way you could with the Apache-2.0 version.

   How to apply: Check Qwen-Image-2.1's actual license terms before shipping it in a product; use the community GGUF build and the official-template-derived ComfyUI workflows to prototype image-editing features quickly.

   image-generationopen-weightslicensing

   Read more: Qwen-Image-2.1 - 7B, top open-weight score on Qwen's own bench, but it's no longer Apache 2.0 · Qwen-Image-2.1-Uncensored-GGUF? · Qwen Image 2.1 best open source Image Editor. Better than Nano Banana - ... · Some very simple Qwen-Image-2.1 workflows

8. #8 ExoRoute: local-first AI protocol gateway and model router in Rusttool

   ExoRoute is a new Rust-written gateway that routes requests across local and cloud model backends behind one stable endpoint.

   Tool · Rust

   One stable endpoint in front of local and cloud model backends

   Routing for cost, latency or privacy — instead of hand-rolled endpoint switching per project.

   Why it matters: Teams juggling several local and hosted models for cost, latency, or privacy reasons need a lightweight router instead of hand-rolling endpoint-switching logic per project.

   How to apply: Evaluate ExoRoute if you want one API surface in front of Ollama/llama.cpp plus cloud fallbacks rather than building custom routing logic yourself.

   local-llmroutinginfrastructure

   Read more: ExoRoute – Local-first AI protocol gateway & model router written in Rust

9. #9 Mini-AGI: continual-learning model architecture trainable on 8GB VRAMrepo

   An open-source continual/online-learning model architecture that trains within an 8GB VRAM budget.

   Why it matters: Continual learning is normally out of reach on consumer hardware; a training approach that fits on a single modest GPU makes on-device incremental-learning experiments accessible to individual engineers.

   How to apply: If you're prototyping a system that needs to adapt from new data over time without full retraining, use the repo as a reference implementation for continual learning on a single consumer GPU.

   trainingopen-sourcelocal-llm

   Read more: Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM

10. #10 Open-source MCP directory adds 1,400+ real-world supplier/factory formsrepo

    A community-maintained MCP directory now indexes 1,400+ scraped supplier and factory forms, including 330 new Korean factory forms this round.

    Why it matters: Most physical-world suppliers don't expose agent-friendly APIs; a maintained MCP directory of scraped forms is a practical bridge for agents doing sourcing or procurement tasks.

    How to apply: If your agent needs to interact with real supplier/vendor forms, check this open-source MCP directory before building a bespoke scraper from scratch.

    mcpagentsopen-source

    Read more: Added ~330 Korean factory forms to my open-source MCP directory (1,400+ total now). What should I scrape next?