Add disk checker CLI

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>

src/main.rs

@@ -0,0 +1,91 @@
+use std::io::{self, Write};
+use std::path::PathBuf;
+use std::process::ExitCode;
+
+use anyhow::Context;
+use clap::Parser;
+use disk_checker::{ScanConfig, parse_byte_count, scan_paths, write_human_report};
+
+#[derive(Debug, Parser)]
+#[command(
+    author,
+    version,
+    about = "Fast folder scanner for sizes, symlinks, hard links, and possible duplicate files",
+    long_about = "disk-checker scans one or more folders, groups files by size, hashes only the first N bytes of same-size candidates, and reports possible duplicates plus symlinks, hard links, special files, and scan errors."
+)]
+struct Cli {
+    /// Folder or file paths to scan. Defaults to the current directory.
+    #[arg(value_name = "PATH")]
+    paths: Vec<PathBuf>,
+
+    /// Number of bytes to hash from each same-size candidate file.
+    /// Accepts plain bytes or units like 512KiB, 1MiB, 2MB.
+    #[arg(long, default_value = "1MiB", value_parser = parse_byte_count)]
+    hash_bytes: u64,
+
+    /// Follow symlinks while scanning. Symlinks are still reported separately.
+    #[arg(long)]
+    follow_links: bool,
+
+    /// Fully hash possible duplicate groups after the fast partial-hash pass.
+    #[arg(long)]
+    verify_full: bool,
+
+    /// Number of worker threads used for hashing. Defaults to Rayon automatic sizing.
+    #[arg(long, value_parser = parse_thread_count)]
+    threads: Option<usize>,
+
+    /// Print machine-readable JSON instead of the human summary.
+    #[arg(long)]
+    json: bool,
+}
+
+fn parse_thread_count(input: &str) -> Result<usize, String> {
+    let threads = input
+        .parse::<usize>()
+        .map_err(|error| format!("invalid thread count {input:?}: {error}"))?;
+    if threads == 0 {
+        Err("thread count must be greater than zero".to_string())
+    } else {
+        Ok(threads)
+    }
+}
+
+fn main() -> anyhow::Result<ExitCode> {
+    let cli = Cli::parse();
+
+    if let Some(threads) = cli.threads {
+        rayon::ThreadPoolBuilder::new()
+            .num_threads(threads)
+            .build_global()
+            .context("failed to initialize hashing thread pool")?;
+    }
+
+    let paths = if cli.paths.is_empty() {
+        vec![PathBuf::from(".")]
+    } else {
+        cli.paths
+    };
+
+    let report = scan_paths(ScanConfig {
+        paths,
+        hash_bytes: cli.hash_bytes,
+        follow_links: cli.follow_links,
+        verify_full: cli.verify_full,
+    });
+
+    let stdout = io::stdout();
+    let mut out = stdout.lock();
+    if cli.json {
+        serde_json::to_writer_pretty(&mut out, &report).context("failed to write JSON report")?;
+        writeln!(out).context("failed to finish JSON report")?;
+    } else {
+        write_human_report(&mut out, &report).context("failed to write report")?;
+    }
+
+    if report.summary.errors > 0 {
+        Ok(ExitCode::from(2))
+    } else {
+        Ok(ExitCode::SUCCESS)
+    }
+}