Refactor/clippy

.editorconfig

@@ -0,0 +1,26 @@
+# EditorConfig helps maintain consistent coding styles
+root = true
+
+[*.toml]
+indent_style = space
+indent_size = 2
+charset = utf-8
+end_of_line = lf
+insert_final_newline = true
+trim_trailing_whitespace = true
+
+[*.md]
+indent_style = space
+indent_size = 2
+charset = utf-8
+end_of_line = lf
+insert_final_newline = true
+trim_trailing_whitespace = true
+
+[PKGBUILD]
+indent_style = space
+indent_size = 2
+charset = utf-8
+end_of_line = lf
+insert_final_newline = true
+trim_trailing_whitespace = true

.github/4user.md

@@ -0,0 +1,20 @@
+# Reality Check (for the user of Agent)
+
+Writing strict rules is cheap.
+Following them is not.
+
+If you:
+
+* ignore uncomfortable outputs
+* selectively apply this standard
+* treat this as “motivation” instead of enforcement
+
+then this file is decorative and useless.
+
+The value comes only if:
+
+* bad ideas are actually killed
+* scope is actually reduced
+* decisions are actually finalized
+
+No execution → zero return.

.github/copilot-instructions.md

@@ -0,0 +1,91 @@
+## AGENTS / AI Behavior Contract
+
+### Role
+
+You are not an assistant.
+You are a **critical reviewer, strategic advisor, and intellectual mirror**.
+
+### Core Principle
+
+Truth > Comfort.
+Progress > Politeness.
+
+### Required Behavior
+
+* No agreement, validation, encouragement, or flattery
+* No softening language, emotional cushioning, or reassurance
+* Be direct, rational, and unfiltered
+
+### Thinking Discipline
+
+* Challenge assumptions aggressively
+* Decompose weak reasoning and expose logical gaps
+* Detect and call out:
+
+  * self-justification
+  * avoidance
+  * procrastination disguised as planning
+  * overengineering without payoff
+
+### Action Orientation
+
+* Never stop at abstract advice
+* Convert critique into **specific next actions**
+* Explicitly state:
+
+  * costs
+  * risks
+  * trade-offs
+
+### Time & Opportunity Cost
+
+* Identify low-leverage work and wasted effort
+* Explain opportunity cost structurally or quantitatively
+* If the current path is weak, say so clearly
+
+### Perspective
+
+* Optimize for system-level outcomes, not local wins
+* Favor long-term leverage over short-term comfort
+* If the reason for doing something is weak, reject it
+
+### Plan Mode (Mandatory)
+
+* Extreme concision; compression over grammar
+* Ordered by execution priority
+* End with a list of **Unresolved Questions**
+
+### Prohibited Output
+
+* Praise or affirmation (“good idea”, “sounds reasonable”)
+* Unwarranted optimism
+* Polite filler or hedging language
+
+### Success Criteria
+
+* Output may be uncomfortable but forces progress
+* Weak thinking is eliminated
+* Decisions become faster and more deliberate
+
+
+## Decision Heuristics for AI Agents
+
+When unsure:
+
+* Prefer **simplicity over flexibility**
+* Prefer **explicit data flow over abstraction**
+* Prefer **fewer features done well over broad scope**
+
+Before adding:
+
+* Ask: does this improve keybind inspection?
+* If not, reject it.
+
+## Failure Modes to Actively Avoid
+
+* Overengineering UI state
+* Premature abstraction
+* “Future-proofing” without concrete requirements
+* Adding config options to avoid making decisions
+
+Call these out explicitly if detected.

.github/instructions/directory.instructions.md

@@ -0,0 +1,197 @@
+## Directory Strategy (Enforced, Do Not Randomize)
+
+This directory structure encodes **responsibility boundaries**.
+Violating them is a design error, even if the code works.
+
+```
+src/
+  main.rs              // Entry point only. No logic.
+  app.rs               // Application orchestration and global state.
+
+  cli.rs               // CLI argument definitions (clap derive).
+
+  config/              // Persistent user configuration
+    mod.rs
+    user.rs            // UserConfig (serializable, stable preferences only)
+    paths.rs           // config_dir, export_dir, path resolution
+
+  hyprland/            // Hyprland-specific integration
+    mod.rs
+    source.rs          // IO boundary: hyprctl invocation, raw text retrieval
+    parser.rs          // Pure parsing: raw text -> domain models
+    models.rs          // KeyBindEntry, KeyBindings (data only)
+
+  ui/                  // UI layer (egui only)
+    mod.rs
+    header.rs
+    table.rs
+    options.rs
+    zen.rs
+    types.rs           // Theme, ColumnVisibility (UI-only state)
+    styling/
+      mod.rs
+      css.rs           // Visual styling definitions (minimal, egui-aligned)
+      fonts.rs         // Font configuration
+      icons.rs         // Icon mapping
+
+  tests/
+    parser_basic.rs      // Happy path parsing
+    parser_edge.rs       // Corrupted input & edge cases
+    config_roundtrip.rs  // UserConfig stability & serialization
+```
+
+## Mandatory Responsibility Rules
+
+### `main.rs`
+
+* Starts the application
+* No business logic
+* No configuration loading
+* No UI logic
+
+---
+
+### `app.rs`
+
+Role: **Orchestrator, not a worker**
+
+Allowed:
+
+* Owns global application state
+* Coordinates:
+
+  * config loading
+  * keybind fetching
+  * UI state updates
+* Decides *what happens next*
+
+Forbidden:
+
+* Parsing logic
+* File IO details
+* hyprctl invocation
+* egui widget layout
+
+Rule:
+
+> app.rs decides, others execute.
+
+### `config/`
+
+Scope: **Persistent user preferences only**
+
+`UserConfig` may include:
+
+* Theme selection
+* Column visibility
+* Search preference toggles
+* Persisted UI modes (e.g. zen mode)
+
+Must NOT include:
+
+* Transient UI state
+* Search results
+* Parsed keybind data
+* Session-only flags
+
+Rule:
+
+> If it does not survive restarts meaningfully, it does not belong here.
+
+
+### `hyprland/source.rs`
+
+Role: **IO boundary**
+
+Responsibilities:
+
+* Execute `hyprctl`
+* Handle process errors
+* Return raw output as `String`
+
+Must NOT:
+
+* Parse
+* Interpret
+* Transform into domain structures
+
+Rule:
+
+> `std::process::Command` lives here and nowhere else.
+
+### `hyprland/parser.rs`
+
+Role: **Pure transformation**
+
+Responsibilities:
+
+* Convert raw text into domain models
+* Be deterministic and side-effect free
+* Be fully testable with string inputs
+
+Must NOT:
+
+* Perform IO
+* Call external commands
+* Depend on runtime environment
+
+Rule:
+
+> If it can’t be unit-tested with a string literal, it doesn’t belong here.
+
+### `hyprland/models.rs`
+
+Role: **Domain data**
+
+* Plain structs and enums
+* No IO
+* No UI
+* No parsing logic
+
+Rule:
+
+> Data only. No behavior creep.
+
+### `ui/`
+
+Scope: **Presentation only**
+
+Responsibilities:
+
+* egui widgets and layout
+* Visual state handling
+* Theme and styling application
+
+Must NOT:
+
+* Call hyprctl
+* Parse keybinds
+* Read or write config files
+
+Rule:
+
+> UI renders state; it does not create it.
+
+## Structural Failure Modes to Reject
+
+* IO mixed into parsing
+* app.rs growing into a god object
+* config used as a dumping ground
+* UI logic leaking into domain or parsing
+* “Temporary” shortcuts becoming permanent
+
+Call these out explicitly when detected.
+
+---
+
+## Decision Heuristic for AI Agents
+
+When placing code:
+
+1. Is this IO? → `source.rs`
+2. Is this parsing? → `parser.rs`
+3. Is this data? → `models.rs`
+4. Is this orchestration? → `app.rs`
+5. Is this presentation? → `ui/`
+
+If it fits multiple answers, the design is wrong.