Fix hooks reference: correct field names and PostToolUse output format

assets/prompts/skills/agent-customization/references/hooks.md

@@ -52,23 +52,108 @@ Each hook command supports:
 
 Hooks receive JSON on stdin and can return JSON on stdout.
 
-- Common output: `continue`, `stopReason`, `systemMessage`
-- `PreToolUse` permissions are read from `hookSpecificOutput.permissionDecision` (`allow` | `ask` | `deny`)
-- `PostToolUse` output can block further processing with `decision: block`
+**Important**: All hook input field names use **snake_case** (e.g., `tool_name`, `tool_input`, `tool_use_id`), not camelCase.
 
-`PreToolUse` example output:
+### Common Input Fields
 
+Tool-based hooks (`PreToolUse`, `PostToolUse`, `PostToolUseFailure`, `PermissionRequest`) receive:
+- `tool_name` (string) - The name of the tool being invoked
+- `tool_input` (object) - The input parameters for the tool
+- `tool_use_id` (string) - Unique identifier for this tool invocation
+- `cwd` (string) - Current working directory
+
+`PostToolUse` additionally receives:
+- `tool_response` (string) - The output from the tool execution
+
+`PostToolUseFailure` receives `error` and `is_interrupt` instead of `tool_response`.
+
+### Common Output Fields
+
+All hooks can return:
+- `continue` (boolean) - Whether to continue execution (default: true)
+- `stopReason` (string) - Reason for stopping (if continue is false)
+- `hookSpecificOutput` (object) - Hook-specific output fields (see below)
+
+### Hook-Specific Output
+
+**PreToolUse** hook returns permissions via `hookSpecificOutput`:
 ```json
 {
   "hookSpecificOutput": {
     "hookEventName": "PreToolUse",
     "permissionDecision": "ask",
-    "permissionDecisionReason": "Needs user confirmation"
+    "permissionDecisionReason": "Needs user confirmation",
+    "updatedInput": { "path": "/safe/path" },
+    "additionalContext": "Remember to validate the path"
+  }
+}
+```
+
+**PostToolUse** hook returns additional context via `hookSpecificOutput`:
+```json
+{
+  "hookSpecificOutput": {
+    "hookEventName": "PostToolUse",
+    "additionalContext": "Tool completed successfully. Next steps: verify the output."
   }
 }
 ```
 
-Exit codes:
+### Complete Examples
+
+**PreToolUse Input Example:**
+```json
+{
+  "tool_name": "Bash",
+  "tool_input": { "command": "npm install", "mode": "sync" },
+  "tool_use_id": "toolu_01abc123",
+  "cwd": "/workspace/project"
+}
+```
+
+**PostToolUse Input Example:**
+```json
+{
+  "tool_name": "Edit",
+  "tool_input": { "file_path": "/workspace/src/app.ts", "old_str": "const x = 1;", "new_str": "const x = 2;" },
+  "tool_use_id": "toolu_01xyz789",
+  "tool_response": "File /workspace/src/app.ts updated successfully",
+  "cwd": "/workspace/project"
+}
+```
+
+### Common Tool Names
+
+When filtering tool-based hooks by `tool_name`, use these values:
+
+**File Operations:**
+- `Edit` - Edit existing files
+- `MultiEdit` - Edit multiple files
+- `Write` - Create new files
+- `Read` - Read file contents
+- `NotebookEdit` - Edit Jupyter notebooks
+
+**Code Search & Navigation:**
+- `Grep` - Search file contents
+- `Glob` - Find files by pattern
+- `LS` - List directory contents
+
+**Execution:**
+- `Bash` - Run shell commands
+- `BashOutput` - Get shell command output
+- `KillBash` - Stop shell command
+- `Task` - Run sub-agent task
+
+**Special:**
+- `EnterPlanMode` - Enter planning mode
+- `ExitPlanMode` - Exit planning mode
+- `TodoWrite` - Write to-do items
+- `WebFetch` - Fetch web pages
+- `WebSearch` - Search the web
+- `AskUserQuestion` - Ask the user a question
+
+### Exit Codes
+
 - `0` success
 - `2` blocking error
 - Other values produce non-blocking warnings