refactor: update llm model configurations and add AI agent guidelines

- 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

modules/home/programs/terminal/nvim/config/lua/llm-config.lua

@@ -1,10 +1,6 @@
 local llm_endpoint = "https://llm-api.va.reichard.io"
-local llm_assistant_model = "qwen3-coder-30b-instruct"
-local llm_infill_model = "qwen3-coder-30b-instruct"
-
--- local llm_assistant_model = "devstral-small-2-instruct"
--- local llm_infill_model = "qwen2.5-coder-3b-instruct"
-
+local llm_assistant_model = "qwen3-coder-next-80b-instruct"
+local llm_infill_model = llm_assistant_model
 
 local current_fim = "copilot" -- change this to switch default
 

modules/home/programs/terminal/pi/config/AGENTS.md

@@ -0,0 +1,33 @@
+# AI Agent Guidelines
+
+## Important Rules
+
+1. **Timeout for bash tool**: The `bash` tool MUST have a timeout specified. Without a timeout, the tool will hang indefinitely and cause the task to fail.
+
+2. **File writing**: Do NOT use `cat` with heredocs to write files. Use the `write` tool instead (or `edit` for modifications).
+
+## Example of Correct Usage
+
+### Incorrect (will hang):
+
+```bash
+bash(command="some long-running command")
+```
+
+### Correct (with timeout):
+
+```bash
+bash(command="some command", timeout=30)
+```
+
+### Incorrect (file writing):
+
+```bash
+bash(command="cat > file.txt << 'EOF'\ncontent\nEOF")
+```
+
+### Correct (file writing):
+
+```bash
+write(path="file.txt", content="content")
+```

modules/home/programs/terminal/pi/default.nix

@@ -37,6 +37,9 @@ in
           };
         };
       };
+      ".pi/agent/AGENTS.md" = {
+        source = ./config/AGENTS.md;
+      };
       ".pi/agent/skills" = {
         source = ./config/skills;
         recursive = true;