- Update nvim to use qwen3-coder-next-80b-instruct model
- Add AGENTS.md with AI agent best practices for timeout and file writing
- Update pi config to include agent guidelines
- Refactor llama-swap: remove old models, update quantizations, add tensor splits,
remove GGML_CUDA_ENABLE_UNIFIED_MEMORY flags, and simplify configuration
Change opencode and pi model filtering to use 'coding' type instead of
more generic 'text-generation' type. Update llama-swap model configs to
include 'coding' in metadata type list for relevant models (deepseek-coder,
qwen-coder, mistral, codellama, llama3-8b-instruct-q5).
- Add pi module for terminal configuration
- Include pi-coding-agent package in home.packages
- Configure pi with llama-swap provider via models.json
- Enable pi for mac-va-mbp-work profile
- Added pi-coding-agent package definition with buildNpmPackage
- Includes npm dependencies and native/build inputs for GUI libraries
- Skips generate-models step during build (models already in repo)
- Creates 'pi' executable pointing to coding-agent dist/cli.js
Fixes build issues where generate-models would fail during build phase
- Update opencode from v1.1.48 to v1.1.51
- Refactor build dependencies: remove fzf, add sysctl conditionally for Darwin
- Fix bun to use nixpkgs-unstable input instead of direct dependency
- Add platform-specific outputHash for Darwin and Linux
- Move wrapProgram from postFixup to installPhase
- Remove unnecessary build patch and postPatch/postBuild workarounds
- Fix config path to use absolute home directory
- Add git and home-manager to default shell inputs
- Minor README documentation reordering
Updated llama-swap to version 189 with corresponding hash updates.
Adjusted UI sourceRoot from "ui" to "ui-svelte" to match upstream
directory structure changes.
Add new model "Qwen3 Coder Next (80B) - Instruct" with 262144 context window
and optimized parameters for coding tasks. Uses CUDA unified memory support.
- Update llama-cpp from b7867 to b7898
- Update opencode from v1.1.12 to v1.1.48 with improved build process:
- Replace custom bundle script with official script/build.ts
- Add shell completion support
- Add version check testing
- Simplify node_modules handling
- Update llama-swap service config with new llama.cpp options
- Clarify opencode agent testing workflow in developer and reviewer configs
- Update llama.cpp from b7789 to b7867
- Update llama-swap from v182 to v186
- Add OpenCode conventional commit command configuration
- Add moonshotai Kimi-K2.5 model to llama-swap
- Add new `qwen3-8b-vision` model with multimodal support using mmproj file
- Add new `qwen2.5-coder-7b-instruct` model with FIM enabled via `--fim-qwen-7b-default`
- Update CUDA device usage from `CUDA0` to `CUDA1` for `olmoe-7b-instruct` and `phi-mini-8b-instruct`
- Upgrade llama.cpp to version 7426 with updated hash and CUDA architectures (61;86)
- Add Copilot acceptance shortcut `<C-J>` in insert mode and disable tab mapping
- Improve cache type settings across multiple models for better performance
This commit includes significant updates to the Nix configuration including:
- Added common NixOS module with system packages
- Updated home-manager configurations for various systems
- Enhanced terminal programs including Claude Code integration
- Updated LLM server configurations with new models
- Improved btop configuration with custom package
- Added opencode support with permission settings
- Updated tmux and nvim configurations
The changes introduce new features, update existing configurations, and improve the overall system setup.
Upgrade the NixOS `stateVersion` from 25.05 to 25.11 in all system configuration files and update all nix channel URLs to the 25.11 releases. Adjust Home Manager and terminal module configurations accordingly, clean up duplicated SSH keys, and update LLaMA server command lines. Minor housekeeping changes to package definitions and other configuration files are also included.
