InFileNameWeTrust

InFileNameWeTrust is a whimsical experiment that stores entire file contents in the filenames themselves. This unconventional method is the ultimate proof-of-concept for "hidden" storage via filename abuse on Windows.

Each file is 0 bytes, yet it secretly embeds all your data in its name. It’s silly, it’s impractical, but it works—and that’s precisely why we love it.

No disclaimers, just pure madness. InFileNameWeTrust!

The project supports both local file system storage and optional integration with Azure Blob Storage.

How It Works

1. Compression
   The input file is compressed with zlib.

2. Base‑N Encoding
   The compressed bytes are interpreted as a large integer and converted into a high‑base representation using a custom valid Unicode BMP alphabet allowed in Windows filenames.

3. Chunking
   Since Windows limits filename length (usually 255 UTF‑16 code units), the encoded string is split into chunks small enough to serve as filenames. Each chunk becomes a zero‑byte file whose name encodes part of your data.

4. Decoding

   • The filenames are sorted by their numeric prefix.
   • The encoded chunks are concatenated.
   • The data is decoded from base‑N back to bytes.
   • zlib decompression restores the original file.

Installation

Prerequisites

• Python 3.10 or newer
• Poetry for dependency management

Steps

1. Clone the Repository

   git clone https://github.com/Nicolas-Prevot/InFileNameWeTrust.git
   cd InFileNameWeTrust

2. Create the Environment

   If you use Conda:

   conda env create -f environment.yml
   conda activate infilenamewetrust

3. Install with Poetry

   poetry install

4. Optional: Install Azure Dependencies To enable Azure Blob Storage functionality, install the optional dependencies:

   poetry install --extras "azure"

Usage

Local Storage

Encoding

Store a file’s data in zero-byte filenames on the local file system:

poetry run encode /path/to/input_file /path/to/output_dir --chunk_size 190 --segment_size 300000

• input_file: The file you want to store in filenames.
• output_dir: The directory to create the "chunk files".
• Options:
  • --chunk_size <N>: Max BMP characters per chunk (default=190).
  • --segment_size <N>: Compressed segment size in bytes (default: 300000).

Decoding

Reconstruct the original file from the chunked filenames:

poetry run decode /path/to/output_dir /path/to/restored_file

• input_dir: The directory containing the chunk files.
• output_file: Where to write the reconstructed data.

Example

poetry run encode README.md ./data/README_md
poetry run decode ./data/README_md ./data/README.md

Azure Blob Storage Usage

The project includes support for storing and retrieving encoded segments as Azure blobs.

Encoding Azure

poetry run azure_encode /path/to/input_file container_name blob_prefix --chunk_size 3650 --segment_size 6_200_000

• input_file: The file you want to store in filenames.
• container_name: Name of the Azure Blob container.
• blob_prefix: Blob "folder" or prefix to use for storage.
• Options:
  • --chunk_size <N>: Max BMP characters per chunk (default=3650).
  • --segment_size <N>: Compressed segment size in bytes (default: 6200000).

Decoding Azure

poetry run azure_decode container_name blob_prefix /path/to/restored_file

• container_name: Azure Blob container name.
• blob_prefix: Blob prefix where chunks are stored.
• output_file: Local file path for the decoded file.

Example Azure

poetry run azure_encode README.md mycontainer myblob/README_md
poetry run azure_decode mycontainer myblob/README_md ./data/README.md

Build the Cython Module (Optional)

If you wish to rebuild the Cython module:

1. Navigate to the source directory:

   cd src/infilenamewetrust

2. Build the extension in place:

   python setup.py build_ext --inplace

Fancy Ideas to Explore

• Alternate Data Streams (ADS): Hide data in filename:stream to store "secret" streams.
• Timestamps: Encode small blocks of data in the creation/mod/access times.
• Short (8.3) Names: If 8.3 short-name generation is enabled, store data in the DOS short name.
• Folder Names / ACLs / Reparse Points: Exploit other NTFS features as "containers" for data.
• Registry Keys: Similar concept, but storing chunked data in registry subkeys or value names.