diff --git a/claude-code/README.md b/claude-code/README.md
deleted file mode 100644
index 87ca28d..0000000
--- a/claude-code/README.md
+++ /dev/null
@@ -1,21 +0,0 @@
-# Srid's Claude Code Configuration
-
-## Nix module
-
-- `nix/home-manager-module.nix` - Home-manager module for auto-wiring this directory layout.
-
-### Usage
-
-Import the home-manager module and set `autoWire.dir`:
-
-```nix
-{
-  imports = [ ./claude-code/nix/home-manager-module.nix ];
-
-  programs.claude-code = {
-    enable = true;
-    autoWire.dir = ./claude-code;
-  };
-}
-```
-
diff --git a/claude-code/commands/hack.md b/claude-code/commands/hack.md
deleted file mode 100644
index ada83fd..0000000
--- a/claude-code/commands/hack.md
+++ /dev/null
@@ -1,54 +0,0 @@
----
-name: hack
-description: Implement a GitHub issue end-to-end until CI passes
-args:
-  - name: issue_url
-    description: GitHub issue URL (e.g., https://github.com/user/repo/issues/123)
-    required: true
-  - name: plan_mode
-    description: Whether to run in plan mode (true/false, default false)
-    required: false
----
-
-This command implements a GitHub issue from start to finish, ensuring all CI checks pass.
-
-**Usage**: `/hack <github-issue-url> [plan_mode]`
-
-**What it does**:
-1. Fetches the GitHub issue details using `gh` CLI
-2. Analyzes the issue requirements and creates an implementation plan
-3. If plan_mode=true, presents plan for approval before implementing
-4. Implements the requested feature or fix
-5. Creates a commit with proper issue reference
-6. Runs `om ci` to validate the implementation
-7. Iterates and fixes any CI failures until all checks pass
-
-**Workflow**:
-1. Parse the GitHub issue URL to extract repository and issue number
-2. Use `gh api` to fetch issue title, description, and labels
-3. Create a detailed implementation plan based on issue requirements
-4. Implement the solution step by step
-5. Commit changes with concise description
-6. Run `om ci` and analyze any failures
-7. Fix CI issues and re-run until all checks pass
-8. Provide final status report
-
-**Prerequisites**:
-- GitHub CLI (`gh`) must be authenticated
-- Must be in a git repository with a non-mainline branch
-- omnix (`om`) must be available for CI checks
-- Repository must have CI configured via omnix
-
-**Error handling**:
-- Validates GitHub URL format
-- Handles GitHub API errors gracefully
-- Provides detailed feedback on CI failures
-- Supports iterative fixing until CI passes
-
-**Examples**:
-```
-/hack https://github.com/myorg/myproject/issues/42
-/hack https://github.com/myorg/myproject/issues/42 true
-```
-
-First example directly implements issue #42. Second example creates a plan first and waits for approval before implementing.
\ No newline at end of file
diff --git a/claude-code/commands/pr.md b/claude-code/commands/pr.md
deleted file mode 100644
index 0017672..0000000
--- a/claude-code/commands/pr.md
+++ /dev/null
@@ -1,65 +0,0 @@
-# PR Command
-
-Create and open a draft pull request on GitHub for the current branch.
-
-## Workflow
-
-1. **Gather Information**
-   - Get the current branch name using `git branch --show-current`
-   - Get the default branch (usually `main` or `master`) using `git remote show origin | grep 'HEAD branch'`
-   - Analyze the diff from the default branch to understand all changes (ignore individual commits)
-   - Review changed files to understand the scope of changes
-
-2. **Generate PR Content**
-   - Create a descriptive title that summarizes the changes
-   - Generate a comprehensive description with:
-     - A brief summary paragraph at the top
-     - **User-Facing Changes**: What end-users will notice or experience
-     - **Developer Notes**: Technical details, implementation notes, or items of interest to other developers
-   - Keep the description concise but informative
-
-3. **Get User Confirmation**
-   - Present the proposed PR title and description to the user
-   - Ask for confirmation or request modifications
-   - Allow the user to edit the title or description before proceeding
-
-4. **Create Draft PR**
-   - Once confirmed, use the `gh` CLI to create a draft PR:
-     ```bash
-     gh pr create --draft --title "TITLE" --body "DESCRIPTION"
-     ```
-   - Report the PR URL to the user
-
-## Requirements
-
-- The `gh` CLI must be installed and authenticated (if not installed, use `nix run nixpkgs#gh` to run it)
-- The repository must have a remote configured on GitHub
-- The current branch must have commits that aren't in the default branch
-
-## Example Output Format
-
-```markdown
-PR Title: Add new authentication system
-
-PR Description:
-This PR introduces a modern OAuth-based authentication system alongside improved password management, enhancing both security and user experience.
-
-## User-Facing Changes
-
-- Users can now log in using OAuth providers (Google, GitHub)
-- Improved password reset flow with better email notifications
-- Session timeout increased from 1 hour to 24 hours
-
-## Developer Notes
-
-- Implemented OAuth2 client using the `oauth2-client` library
-- Added new `AuthService` class to handle authentication logic
-- Updated database schema with new `oauth_tokens` table
-- Added comprehensive unit tests for authentication flows
-```
-
-## Notes
-
-- Always create as a **draft** PR initially so the user can make further changes if needed
-- If the branch name follows a convention (e.g., `feature/`, `fix/`), incorporate that context into the title
-- If there are uncommitted changes, warn the user before proceeding
diff --git a/claude-code/mcp/chrome-devtools.nix b/claude-code/mcp/chrome-devtools.nix
deleted file mode 100644
index 5e6ee8f..0000000
--- a/claude-code/mcp/chrome-devtools.nix
+++ /dev/null
@@ -1,4 +0,0 @@
-{
-  command = "npx";
-  args = [ "chrome-devtools-mcp@latest" ];
-}
diff --git a/claude-code/mcp/nixos-mcp.nix b/claude-code/mcp/nixos-mcp.nix
deleted file mode 100644
index a8c2457..0000000
--- a/claude-code/mcp/nixos-mcp.nix
+++ /dev/null
@@ -1,4 +0,0 @@
-{
-  command = "uvx";
-  args = [ "mcp-nixos" ];
-}
diff --git a/claude-code/memory.md b/claude-code/memory.md
deleted file mode 100644
index 7474f83..0000000
--- a/claude-code/memory.md
+++ /dev/null
@@ -1,12 +0,0 @@
-# System Instructions
-
-- Talk like a naive kid as much as possible
-- Sacrifice grammar for brevity.
-- Don't say "First, " say "Indeed, "
-- Don't say "Good" or "Perfect" say "Great"
-
-# Tools
-
-- **gh**: If `gh` is unavailable, get it from nixpkgs, viz.: `nix run nixpkgs#gh`.
-- **article-extractor**: To extract clean article content from URLs (blog posts, articles, tutorials), use the article-extractor skill. It removes ads, navigation, and clutter and saves readable text.
-
diff --git a/claude-code/nix/home-manager-module.nix b/claude-code/nix/home-manager-module.nix
deleted file mode 100644
index ee4ee13..0000000
--- a/claude-code/nix/home-manager-module.nix
+++ /dev/null
@@ -1,124 +0,0 @@
-{ config, lib, pkgs, ... }:
-
-let
-  cfg = config.programs.claude-code;
-
-  # Auto-wire subagents from subagents/ directory
-  subagentsDir = cfg.autoWire.dir + "/subagents";
-  autoAgents = lib.optionalAttrs (builtins.pathExists subagentsDir)
-    (lib.mapAttrs'
-      (fileName: _:
-        lib.nameValuePair
-          (lib.removeSuffix ".md" fileName)
-          (builtins.readFile (subagentsDir + "/${fileName}"))
-      )
-      (builtins.readDir subagentsDir));
-
-  # Auto-wire commands from commands/ directory
-  commandsDir = cfg.autoWire.dir + "/commands";
-  autoCommands = lib.optionalAttrs (builtins.pathExists commandsDir)
-    (lib.mapAttrs'
-      (fileName: _:
-        lib.nameValuePair
-          (lib.removeSuffix ".md" fileName)
-          (builtins.readFile (commandsDir + "/${fileName}"))
-      )
-      (builtins.readDir commandsDir));
-
-  # Auto-wire skills from skills/ directory
-  skillsDir = cfg.autoWire.dir + "/skills";
-  skillDirs = lib.optionalAttrs (builtins.pathExists skillsDir)
-    (lib.filterAttrs (_: type: type == "directory") (builtins.readDir skillsDir));
-
-  # Process skill: if it has default.nix, build and substitute; otherwise use as-is
-  processSkill = skillName:
-    let
-      skillPath = skillsDir + "/${skillName}";
-      hasDefaultNix = builtins.pathExists (skillPath + "/default.nix");
-    in
-    if hasDefaultNix then
-      let
-        skillPkg = pkgs.callPackage skillPath { };
-      in
-      pkgs.runCommand "${skillName}-skill" { } ''
-        mkdir -p $out
-        substitute ${skillPath}/SKILL.md $out/SKILL.md \
-          --replace-fail '@${skillName}@' '${skillPkg}/bin/${skillName}'
-      ''
-    else
-      skillPath;
-
-  # Auto-wire MCP servers from mcp/ directory
-  mcpDir = cfg.autoWire.dir + "/mcp";
-  autoMcpServers = lib.optionalAttrs (builtins.pathExists mcpDir)
-    (lib.mapAttrs'
-      (fileName: _:
-        lib.nameValuePair
-          (lib.removeSuffix ".nix" fileName)
-          (import (mcpDir + "/${fileName}"))
-      )
-      (builtins.readDir mcpDir));
-
-  # Auto-load settings from settings.nix
-  settingsFile = cfg.autoWire.dir + "/settings.nix";
-  autoSettings = lib.optionalAttrs (builtins.pathExists settingsFile)
-    (import settingsFile);
-
-  # Auto-load memory from memory.md
-  memoryFile = cfg.autoWire.dir + "/memory.md";
-  autoMemory = lib.optionalAttrs (builtins.pathExists memoryFile)
-    { text = builtins.readFile memoryFile; };
-
-in
-{
-  options.programs.claude-code = {
-    autoWire = {
-      enable = lib.mkOption {
-        type = lib.types.bool;
-        default = true;
-        description = ''
-          Whether to automatically wire up subagents, commands, skills, and MCP servers from autoWire.dir.
-          Set to false if you want to manually configure these.
-        '';
-      };
-
-      dir = lib.mkOption {
-        type = lib.types.nullOr lib.types.path;
-        default = null;
-        description = ''
-          Path to the claude-code directory containing subagents/, commands/, skills/, mcp/, settings.nix, and memory.md.
-          When set, these will be automatically discovered and configured.
-        '';
-      };
-    };
-  };
-
-  config = lib.mkIf (cfg.enable && cfg.autoWire.dir != null && cfg.autoWire.enable) {
-    # Link skill directories to ~/.claude/skills/
-    home.file = lib.mapAttrs'
-      (skillName: _:
-        lib.nameValuePair ".claude/skills/${skillName}" {
-          source = processSkill skillName;
-          recursive = true;
-        }
-      )
-      skillDirs;
-
-    programs.claude-code = {
-      # Auto-wire settings if settings.nix exists
-      settings = lib.mkDefault autoSettings;
-
-      # Auto-wire memory if memory.md exists
-      memory = lib.mkDefault autoMemory;
-
-      # Auto-wire commands from commands/ directory
-      commands = lib.mkDefault autoCommands;
-
-      # Auto-wire agents from subagents/ directory
-      agents = lib.mkDefault autoAgents;
-
-      # Auto-wire MCP servers from mcp/ directory
-      mcpServers = lib.mkDefault autoMcpServers;
-    };
-  };
-}
diff --git a/claude-code/settings.nix b/claude-code/settings.nix
deleted file mode 100644
index 507e1ed..0000000
--- a/claude-code/settings.nix
+++ /dev/null
@@ -1,8 +0,0 @@
-{
-  # theme = "dark";
-  permissions = {
-    defaultMode = "bypassPermissions";
-  };
-  # Disable Claude from adding itself as co-author to commits
-  includeCoAuthoredBy = false;
-}
diff --git a/claude-code/skills/article-extractor/SKILL.md b/claude-code/skills/article-extractor/SKILL.md
deleted file mode 100644
index 7ad1f28..0000000
--- a/claude-code/skills/article-extractor/SKILL.md
+++ /dev/null
@@ -1,31 +0,0 @@
----
-name: article-extractor
-description: Extract clean article content from URLs using reader. Use when user wants to download/extract/save an article from a URL.
-allowed-tools:
-  - Bash
-  - Write
----
-
-# Article Extractor
-
-Extracts clean article content from URLs using `reader` (Mozilla Readability).
-
-## When to Use
-
-User wants to:
-- Extract article from URL
-- Download blog post content
-- Save article as text
-
-## Workflow
-
-Extract article content and save it:
-
-```bash
-CONTENT=$(@article-extractor@ "$URL")
-```
-
-Then use the Write tool to save `$CONTENT` to the desired location with an appropriate filename.
-
-The script outputs clean markdown content to stdout. You decide where and how to save it based on the user's request.
-
diff --git a/claude-code/skills/article-extractor/default.nix b/claude-code/skills/article-extractor/default.nix
deleted file mode 100644
index d9cf476..0000000
--- a/claude-code/skills/article-extractor/default.nix
+++ /dev/null
@@ -1,23 +0,0 @@
-{ writeShellApplication, curl, reader }:
-
-writeShellApplication {
-  name = "article-extractor";
-
-  runtimeInputs = [ curl reader ];
-
-  text = ''
-    URL="$1"
-
-    # Create unique temp file
-    TEMP_HTML=$(mktemp)
-
-    # Download HTML
-    curl -s -L "$URL" > "$TEMP_HTML"
-
-    # Extract and output article to stdout
-    reader "$TEMP_HTML"
-
-    # Clean up
-    rm "$TEMP_HTML"
-  '';
-}
diff --git a/claude-code/skills/haskell/SKILL.md b/claude-code/skills/haskell/SKILL.md
deleted file mode 100644
index a1130ce..0000000
--- a/claude-code/skills/haskell/SKILL.md
+++ /dev/null
@@ -1,129 +0,0 @@
----
-name: haskell
-description: Expert Haskell development assistance. Use when working with Haskell code, .hs files, Cabal, ghcid, or when user mentions Haskell, functional programming, or type-level programming.
----
-
-# Haskell Development
-
-Expert assistance for Haskell programming.
-
-## Guidelines
-
-**CRITICAL - Error Handling in Code**: NEVER write code that silently ignores errors:
-- Do NOT use `undefined` or `error` as placeholders
-- Do NOT skip handling error cases in pattern matches
-- Do NOT ignore `Maybe`/`Either` failure cases
-- Handle all possible cases explicitly
-- Use types to make impossible states unrepresentable
-
-Every error case in generated code must be handled properly.
-
-**Code Quality**:
-- Write type signatures for all top-level definitions
-- Write total functions (avoid `head`, `tail`)
-- Prefer pure functions over IO when possible
-- Use explicit exports in modules
-- Leverage type system for safety
-- Favor composition over complex functions
-- Write Haddock documentation for public APIs
-
-**Idiomatic Patterns**:
-- Prefer `Text` over `String`
-- Use `newtype` wrappers for domain types
-- Apply smart constructors for validation
-- Use RecordDotSyntax (`value.field`) instead of manually unpacking data types
-- Use lenses for record manipulation when appropriate
-- Use `Applicative` and `Monad` appropriately
-- Avoid trivial `let` bindings when inlining keeps code simple and readable
-
-## HLint
-
-**IMPORTANT**: Always check if the project has a `.hlint.yaml` file. If it does:
-1. Read the `.hlint.yaml` file to understand project-specific style rules
-2. Follow all recommendations in that file when writing code
-3. After making code changes, run `hlint` to ensure no warnings
-
-## Relude Best Practices
-
-When using relude prelude, follow these HLint recommendations (from https://github.com/kowainik/relude/blob/main/.hlint.yaml):
-
-**Basic Idioms**:
-- Use `pass` instead of `pure ()` or `return ()`
-- Use `one` instead of `(: [])`, `(:| [])`, or singleton functions
-- Use `<<$>>` for double fmap: `f <<$>> x` instead of `fmap (fmap f) x`
-- Use `??` (flap) operator: `ff ?? x` instead of `fmap ($ x) ff`
-
-**File I/O**:
-- `readFileText`, `writeFileText`, `appendFileText` for Text files
-- `readFileLText`, `writeFileLText`, `appendFileLText` for lazy Text
-- `readFileBS`, `writeFileBS`, `appendFileBS` for ByteString
-- `readFileLBS`, `writeFileLBS`, `appendFileLBS` for lazy ByteString
-
-**Console Output**:
-- `putText`, `putTextLn` for Text
-- `putLText`, `putLTextLn` for lazy Text
-- `putBS`, `putBSLn` for ByteString
-- `putLBS`, `putLBSLn` for lazy ByteString
-
-**Maybe/Either Helpers**:
-- `whenJust m f` instead of `maybe pass f m`
-- `whenJustM m f` for monadic versions
-- `whenNothing_ m x` / `whenNothingM_ m x` for Nothing cases
-- `whenLeft_ m f`, `whenRight_ m f` for Either
-- `whenLeftM_ m f`, `whenRightM_ m f` for monadic Either
-- `leftToMaybe`, `rightToMaybe` for conversions
-- `maybeToRight l`, `maybeToLeft r` for conversions
-- `isLeft`, `isRight` instead of `either (const True/False) (const False/True)`
-
-**List Operations**:
-- Use `ordNub` instead of `nub` (O(n log n) vs O(n²))
-- Use `sortNub` instead of `Data.Set.toList . Data.Set.fromList`
-- Use `sortWith f` instead of `sortBy (comparing f)` for simple cases
-- Use `viaNonEmpty f x` instead of `fmap f (nonEmpty x)`
-- Use `asumMap f xs` instead of `asum (map f xs)`
-- Use `toList` instead of `foldr (:) []`
-
-**Monadic Operations**:
-- `andM s` instead of `and <$> sequence s`
-- `orM s` instead of `or <$> sequence s`
-- `allM f s` instead of `and <$> mapM f s`
-- `anyM f s` instead of `or <$> mapM f s`
-- `guardM f` instead of `f >>= guard`
-- `infinitely` instead of `forever` (better typed)
-- `unlessM (not <$> x)` → use `whenM x` instead
-- `whenM (not <$> x)` → use `unlessM x` instead
-
-**State/Reader Operations**:
-- `usingReaderT` instead of `flip runReaderT`
-- `usingStateT` instead of `flip runStateT`
-- `evaluatingStateT s st` instead of `fst <$> usingStateT s st`
-- `executingStateT s st` instead of `snd <$> usingStateT s st`
-
-**Transformer Lifting**:
-- `hoistMaybe m` instead of `MaybeT (pure m)`
-- `hoistEither m` instead of `ExceptT (pure m)`
-
-**List Pattern Matching**:
-- `whenNotNull m f` for `case m of [] -> pass; (x:xs) -> f (x :| xs)`
-- `whenNotNullM m f` for monadic version
-
-**Text/ByteString Conversions**:
-- Use relude's `toText`, `toString`, `toLText` instead of pack/unpack
-- Use relude's `encodeUtf8`, `decodeUtf8` for UTF-8 encoding
-- `fromStrict`, `toStrict` for lazy/strict conversions
-
-## Testing
-
-- Use QuickCheck for property-based testing
-- Use HUnit or Hspec for unit tests
-- Provide good examples in documentation
-
-## Build instructions
-
-As you make code changes, start a subagent in parallel to resolve any compile errors in `ghcid.txt`.
-
-**IMPORTANT**: Do not run build commands yourself. The human runs ghcid in the terminal, which then updates `ghcid.txt` with any compile error or warning (if this file does not exist, or if ghcid has stopped, remind the human to address it). You should read `ghcid.txt` (in _entirety_) after making code changes; this file updates near-instantly.
-
-**Adding/Deleting modules**: When a new `.hs` file is added or deleted, the `.cabal` file must be updated accordingly. However, if `package.yaml` exists in the project, run `hpack` instead to regenerate the `.cabal` file with the updated module list. This will trigger `ghcid` to restart automatically.
-
-**HLint warnings**: Once all code changes are made and `ghcid.txt` shows success, check if the project has a `.hlint.yaml` file. If it does, run hlint to ensure there are no warnings and address any that appear.
diff --git a/claude-code/skills/nix/SKILL.md b/claude-code/skills/nix/SKILL.md
deleted file mode 100644
index ff02770..0000000
--- a/claude-code/skills/nix/SKILL.md
+++ /dev/null
@@ -1,11 +0,0 @@
----
-name: nix
-description: Expert Nix development assistance. Use when working with Nix code, .nix files, flakes, NixOS configuration, or when user mentions Nix, flake-parts, nixpkgs.
----
-
-# Nix Development
-
-## Guidelines
-
-- Use **flake-parts** when creating new flakes
-- Format with **nixpkgs-fmt** after changes
diff --git a/claude-code/subagents/pre-commit.md b/claude-code/subagents/pre-commit.md
deleted file mode 100644
index b955600..0000000
--- a/claude-code/subagents/pre-commit.md
+++ /dev/null
@@ -1,31 +0,0 @@
----
-name: pre-commit
-description: Invoke after changing sources locally, and only if git-hooks.nix is used by Nix.
-tools: Bash
----
-# Pre-commit Quality Check Agent
-
-## Purpose
-This agent runs `pre-commit run -a` to automatically check code quality and formatting when other agents modify files in the repository.
-
-## When to Use
-- After any agent makes file modifications
-- Before committing changes
-- When code quality checks are needed
-
-## Tools Available
-- Bash (for running pre-commit)
-- Read (for checking file contents if needed)
-
-## Typical Workflow
-1. Run `pre-commit run -a` to check all files
-2. Report any issues found
-3. Suggest fixes if pre-commit hooks fail
-4. Re-run after fixes are applied
-
-## Example Usage
-```bash
-pre-commit run -a
-```
-
-This agent ensures code quality standards are maintained across the repository by leveraging the configured pre-commit hooks.
\ No newline at end of file
diff --git a/flake.lock b/flake.lock
index 3f00797..a8f26e6 100644
--- a/flake.lock
+++ b/flake.lock
@@ -1,5 +1,21 @@
 {
   "nodes": {
+    "AI": {
+      "flake": false,
+      "locked": {
+        "lastModified": 1762823677,
+        "narHash": "sha256-tAFietBkeOYSUh+5KqZcfDA55L+A9IsPQZSuupNY+vM=",
+        "owner": "srid",
+        "repo": "AI",
+        "rev": "ba61df36131e5a9a7165c141e7e9ba18a56da324",
+        "type": "github"
+      },
+      "original": {
+        "owner": "srid",
+        "repo": "AI",
+        "type": "github"
+      }
+    },
     "agenix": {
       "inputs": {
         "darwin": "darwin",
@@ -895,6 +911,7 @@
     },
     "root": {
       "inputs": {
+        "AI": "AI",
         "agenix": "agenix",
         "disko": "disko",
         "flake-parts": "flake-parts",
diff --git a/flake.nix b/flake.nix
index 0d1a5b7..20c2b0e 100644
--- a/flake.nix
+++ b/flake.nix
@@ -31,6 +31,8 @@
     nix-ai-tools.url = "github:numtide/nix-ai-tools";
     nix-ai-tools.inputs.nixpkgs.follows = "nixpkgs";
     landrun-nix.url = "github:srid/landrun-nix";
+    AI.url = "github:srid/AI";
+    AI.flake = false;
 
     # Neovim
     nixvim.url = "github:nix-community/nixvim";
diff --git a/modules/flake-parts/nixos-flake.nix b/modules/flake-parts/nixos-flake.nix
index 3967663..427b68c 100644
--- a/modules/flake-parts/nixos-flake.nix
+++ b/modules/flake-parts/nixos-flake.nix
@@ -17,6 +17,7 @@
         "nix-index-database"
         "nixvim"
         "nix-ai-tools"
+        "AI"
       ];
     };
   };
diff --git a/modules/home/claude-code/default.nix b/modules/home/claude-code/default.nix
index ae9c768..3df591d 100644
--- a/modules/home/claude-code/default.nix
+++ b/modules/home/claude-code/default.nix
@@ -1,8 +1,11 @@
 { flake, pkgs, ... }:
+let
+  inherit (flake.inputs) AI;
+in
 {
   # Import the general home-manager module
   imports = [
-    ../../../claude-code/nix/home-manager-module.nix
+    "${AI}/nix/home-manager-module.nix"
   ];
 
   # Packages often used by Claude Code CLI
@@ -20,6 +23,6 @@
       else pkgs.claude-code;
 
     # Set the claude-code directory for auto-wiring
-    autoWire.dir = ../../../claude-code;
+    autoWire.dir = AI;
   };
 }