Loading ToolBench
Multi-format steganography toolkit. Hide files in audio and images with AES-256-GCM encryption and Argon2id key derivation.
Detect whether an audio file contains hidden steganographic data. Analyzes an audio file for signs of embedded data. If a password is provided via env var or file, attempts to decrypt and report details. SECURITY: Passwords are resolved server-side — never passed as plaintext. Risk level: READ-ONLY (reads file metadata, no writes)
Check how much secret data an audio file can hide. Calculates the steganographic capacity of a carrier audio file at the specified quality level. Risk level: READ-ONLY (reads file metadata, no writes)
Extract hidden files from an encoded audio file. Reverses the steganography encoding process to recover secret files. If the files were encrypted, the correct password is required. SECURITY: Passwords are resolved server-side from environment variables or encrypted files — never passed as plaintext. Extracted file contents are NEVER returned — only filenames and sizes. Risk level: WRITE (writes extracted files to output_dir)
Hide secret files inside an audio carrier file using steganography. Embeds one or more secret files into an audio carrier using LSB (Least Significant Bit) steganography. Optionally encrypts with AES-256-GCM and Argon2id key derivation. SECURITY: Passwords are resolved server-side from environment variables or encrypted files — never passed as plaintext through the AI model. Set the password in an env var before launching the MCP client: export GHOSTBIT_PASSWORD="your_password" Then pass password_env="GHOSTBIT_PASSWORD" to this tool. Risk level: WRITE (reads secret files, writes output file)
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 11 | <=2025-11-25 | v2 |
Detect whether an image contains hidden steganographic data. Analyzes an image for signs of embedded data. Reports the format, algorithm, and whether encryption was detected. SECURITY: Passwords are resolved server-side — never passed as plaintext. Risk level: READ-ONLY (reads file metadata, no writes)
Check how much secret data an image can hide. Calculates the steganographic capacity of an image file. The algorithm is auto-detected from the image format (LSB, Palette, SVG XML). Risk level: READ-ONLY (reads file metadata, no writes)
Extract hidden files from a stego image. Reverses the steganography encoding to recover secret files from an image. If the files were encrypted, the correct password is required. SECURITY: Passwords are resolved server-side — never passed as plaintext. Extracted file contents are NEVER returned — only filenames and sizes. Risk level: WRITE (writes extracted files to output_dir)
Hide secret files inside an image carrier using steganography. Embeds one or more secret files into a cover image. The algorithm is selected automatically based on the image format (LSB for most formats, Palette for GIF, SVG XML embedding for SVG). Optionally encrypts with AES-256-GCM and Argon2id key derivation. SECURITY: Passwords are resolved server-side from environment variables or encrypted files — never passed as plaintext through the AI model. Set the password in an env var before launching the MCP client: export GHOSTBIT_PASSWORD="your_password" Then pass password_env="GHOSTBIT_PASSWORD" to this tool. Risk level: WRITE (reads secret files, writes output image)