--- phase: 02-core-archiver plan: 02 type: execute wave: 2 depends_on: ["02-01"] files_modified: - src/archive.rs - src/main.rs autonomous: true requirements: [CLI-02, CLI-03] must_haves: truths: - "Running `encrypted_archive pack file1 file2 -o out.bin` produces a valid archive file" - "The output archive file starts with magic bytes 0x00 0xEA 0x72 0x63" - "Running `encrypted_archive inspect out.bin` shows file count, names, and sizes without decryption" - "Running `encrypted_archive unpack out.bin -o outdir/` extracts files identical to originals" - "Each file in the archive has a unique random IV (no IV reuse)" - "HMAC is verified before decryption during unpack (reject tampered files)" - "SHA-256 is verified after decompression during unpack (detect corruption)" - "The output file is not recognized by `file` command (no standard signatures)" - "Already-compressed files (APK) are stored without gzip compression" artifacts: - path: "src/archive.rs" provides: "pack(), unpack(), inspect() orchestration functions" exports: ["pack", "unpack", "inspect"] min_lines: 150 - path: "src/main.rs" provides: "CLI dispatch wiring commands to archive functions" contains: "archive::pack" key_links: - from: "src/archive.rs" to: "src/format.rs" via: "writes header and TOC entries using format module" pattern: "format::write_header|format::write_toc_entry|format::read_header" - from: "src/archive.rs" to: "src/crypto.rs" via: "encrypts/decrypts file data and computes/verifies HMAC" pattern: "crypto::encrypt_data|crypto::decrypt_data|crypto::compute_hmac|crypto::verify_hmac" - from: "src/archive.rs" to: "src/compression.rs" via: "compresses/decompresses file data based on should_compress" pattern: "compression::compress|compression::decompress|compression::should_compress" - from: "src/main.rs" to: "src/archive.rs" via: "dispatches CLI commands to archive functions" pattern: "archive::pack|archive::unpack|archive::inspect" --- Implement the three archive commands (pack, unpack, inspect) and wire them to the CLI, producing a fully functional encrypted archive tool. Purpose: This is the core deliverable of Phase 2 -- the working archiver. Pack produces archives conforming to FORMAT.md, inspect reads metadata without decryption, and unpack extracts files with full HMAC and SHA-256 verification. Output: A working `encrypted_archive` binary with pack, unpack, and inspect commands that produce and consume archives matching the FORMAT.md specification. @/home/nick/.claude/get-shit-done/workflows/execute-plan.md @/home/nick/.claude/get-shit-done/templates/summary.md @.planning/PROJECT.md @.planning/ROADMAP.md @.planning/STATE.md @.planning/phases/02-core-archiver/02-RESEARCH.md @.planning/phases/02-core-archiver/02-01-SUMMARY.md @docs/FORMAT.md Task 1: Implement pack and inspect commands src/archive.rs, src/main.rs 1. Implement `pack()` in `src/archive.rs`: Signature: `pub fn pack(files: &[PathBuf], output: &Path, no_compress: &[String]) -> anyhow::Result<()>` Algorithm (two-pass archive writing per research Pattern 2): **Pass 1 -- Process all files:** For each input file: a. Read entire file into memory (per Out of Scope: no streaming). b. Validate file size <= u32::MAX. If exceeded, return error: "File too large: {name} ({size} bytes exceeds 4 GB limit)". c. Compute SHA-256 of original data: `crypto::sha256_hash(&data)`. d. Determine compression: `compression::should_compress(&filename, no_compress)`. e. If compressing: `compressed = compression::compress(&data)?`. Else: `compressed = data.clone()`, set compression_flag = 0. f. Generate random IV: `crypto::generate_iv()`. g. Encrypt: `ciphertext = crypto::encrypt_data(&compressed, &KEY, &iv)`. h. Compute HMAC: `hmac = crypto::compute_hmac(&KEY, &iv, &ciphertext)`. i. Store results in a `ProcessedFile` struct (or equivalent) with: name, original_size, compressed_size, encrypted_size, iv, hmac, sha256, compression_flag, ciphertext. **Pass 2 -- Compute offsets and write archive:** a. Compute TOC size: sum of (101 + name.len()) for each file. b. Set toc_offset = HEADER_SIZE (40). c. Compute data offsets: - First file: data_offset = toc_offset + toc_size - Each subsequent: previous data_offset + previous encrypted_size d. Determine flags byte: if ANY file has compression_flag == 1, set bit 0 (0x01). Bits 1-3 are 0 (no obfuscation in Phase 2). Bits 4-7 MUST be 0. e. Create Header: version=1, flags, file_count, toc_offset, toc_size, toc_iv=[0u8;16], reserved=[0u8;8]. f. Open output file for writing. g. Write header using `format::write_header()`. h. Write TOC entries using `format::write_toc_entry()` for each file (with computed data_offset). i. Write data blocks: for each file, write ciphertext bytes. j. Print summary: "Packed {N} files into {output} ({total_bytes} bytes)". Import KEY from `crate::key::KEY`. 2. Implement `inspect()` in `src/archive.rs`: Signature: `pub fn inspect(archive: &Path) -> anyhow::Result<()>` Algorithm: a. Open archive file for reading. b. Read header using `format::read_header()`. This validates magic, version, flags. c. Read TOC entries using `format::read_toc()`. d. Print header info: ``` Archive: {filename} Version: {version} Flags: 0x{flags:02X} Files: {file_count} TOC offset: {toc_offset} TOC size: {toc_size} ``` e. For each file entry, print: ``` [{i}] {name} Original: {original_size} bytes Compressed: {compressed_size} bytes Encrypted: {encrypted_size} bytes Offset: {data_offset} Compression: {yes/no} IV: {hex} HMAC: {hex} SHA-256: {hex} ``` f. Print total: "Total original size: {sum} bytes". 3. Implement `unpack()` in `src/archive.rs`: Signature: `pub fn unpack(archive: &Path, output_dir: &Path) -> anyhow::Result<()>` Algorithm (follows FORMAT.md Section 10 decode order): a. Open archive file for reading. b. Read header using `format::read_header()`. Validates magic, version, flags. c. Read TOC entries using `format::read_toc()`. d. Create output directory if it doesn't exist. e. Track error count for per-file failures. f. For each file entry: - Seek to data_offset. Read encrypted_size bytes (ciphertext). - **Verify HMAC FIRST**: `crypto::verify_hmac(&KEY, &entry.iv, &ciphertext, &entry.hmac)`. If HMAC fails: print error "HMAC verification failed for {name}, skipping", increment error count, continue to next file. - Decrypt: `decrypted = crypto::decrypt_data(&ciphertext, &KEY, &entry.iv)?`. - Decompress if compression_flag == 1: `decompressed = compression::decompress(&decrypted)?`. Else: `decompressed = decrypted`. - **Verify SHA-256**: Compare `crypto::sha256_hash(&decompressed)` with `entry.sha256`. If mismatch: print warning "SHA-256 mismatch for {name} (data may be corrupted)", increment error count. Still write the file. - Create parent directories if name contains `/`. - Write decompressed data to `output_dir/entry.name`. - Print "Extracted: {name} ({original_size} bytes)". g. Print summary: "Extracted {success_count}/{file_count} files". h. If error_count > 0: return Err with message "{error_count} file(s) had verification errors". Per research open question #2: Abort on header/TOC errors. For per-file errors (HMAC/SHA-256), report and continue with non-zero exit code. 4. Wire up `src/main.rs`: Replace placeholder functions with actual calls: ```rust match cli.command { Commands::Pack { files, output, no_compress } => { archive::pack(&files, &output, &no_compress)?; } Commands::Unpack { archive, output_dir } => { archive::unpack(&archive, &output_dir)?; } Commands::Inspect { archive } => { archive::inspect(&archive)?; } } ``` IMPORTANT: Use `std::io::Seek` and `SeekFrom::Start(offset as u64)` for seeking to data_offset during unpack. IMPORTANT: Use `std::fs::create_dir_all()` for creating output directories. IMPORTANT: filename from TocEntry may contain path separators -- sanitize to prevent directory traversal (reject names starting with `/` or containing `..`). cd /home/nick/Projects/Rust/encrypted_archive && cargo build 2>&1 && echo "--- Build OK ---" && echo "Hello, World!" > /tmp/ea_test_hello.txt && printf '\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10' > /tmp/ea_test_binary.bin && cargo run -- pack /tmp/ea_test_hello.txt /tmp/ea_test_binary.bin -o /tmp/ea_test_archive.bin 2>&1 && echo "--- Pack OK ---" && cargo run -- inspect /tmp/ea_test_archive.bin 2>&1 && echo "--- Inspect OK ---" && mkdir -p /tmp/ea_test_out && cargo run -- unpack /tmp/ea_test_archive.bin -o /tmp/ea_test_out 2>&1 && echo "--- Unpack OK ---" && diff /tmp/ea_test_hello.txt /tmp/ea_test_out/ea_test_hello.txt && diff /tmp/ea_test_binary.bin /tmp/ea_test_out/ea_test_binary.bin && echo "--- Round-trip OK: files identical ---" && file /tmp/ea_test_archive.bin 2>&1 && echo "--- file command check (should say 'data' not a known format) ---" && rm -rf /tmp/ea_test_* Verify inspect output shows correct file metadata, verify round-trip produces identical files run after this task commits, before declaring plan complete - `encrypted_archive pack` accepts multiple input files and produces a single archive - `encrypted_archive inspect` displays file metadata (names, sizes, offsets, IVs, HMACs, SHA-256) without decryption - `encrypted_archive unpack` extracts all files, verifying HMAC before decryption and SHA-256 after decompression - Round-trip test: pack 2 files, unpack, diff shows files are byte-identical - Archive file is not recognized by `file` command (shows "data" not a known format) - HMAC verification happens before decryption (encrypt-then-MAC correctly implemented) - Files with compressed extensions (apk, zip, etc.) are stored without gzip compression - Error handling: HMAC failure skips file and continues, SHA-256 mismatch warns but still writes - `cargo build` succeeds - Pack 2 test files (text + binary), unpack to different directory, diff shows byte-identical - `file archive.bin` does NOT show a recognized format (should show "data") - `inspect` shows correct file count, names, sizes, and hex-encoded IVs/HMACs/SHA-256 - Compression auto-detection: `.apk` file should have compression_flag=0 in inspect output - Unpack with tampered archive (flip a byte in ciphertext) should report HMAC failure but continue A fully functional `encrypted_archive` binary where `pack` creates archives matching FORMAT.md, `inspect` displays metadata, and `unpack` extracts with full HMAC and SHA-256 verification. Round-trip fidelity: packed files are byte-identical after unpacking. After completion, create `.planning/phases/02-core-archiver/02-02-SUMMARY.md`