document-security

Local privacy preflight for files, folders, and archives—catch secrets, PII, metadata, and hidden document content before sharing.

ScriptoX is a versatile tool that enables users to create SVG and PDF files with embedded JavaScript code.

Official implementation of VIF: A novel constructive watermarking system for document security and forensic tracking using Morphological Dilation and AI-OCR. [Patent Pending - SAIP]

Detects hidden text and prompt injection in PDF and DOCX documents: white-on-white text, sub-legible fonts, off-page text, invisible render mode and invisible Unicode characters.

Fail-closed document inspection before untrusted files reach an LLM

Banking-optimized signature verification with Siamese CNN - Contrastive learning for forgery detection with high recall for fraud prevention

Privacy-first Windows app that finds sensitive data before files are shared. Scans locally, supports optional on-device Ollama AI, and creates redacted copies without uploading source files.

Unified scanner for source code and documents — rule-based vulnerability and malware detection

Redact PDFs on your own machine. Content is removed from the file, not hidden under a black rectangle, and a post-redaction audit blocks any export that still contains it.

Permanently redact sensitive text from PDF documents — native macOS app with irrecoverable text removal

Vet the applicants. Employer-side resume screening: scores each CV against your JD and flags hidden text, prompt injection, keyword stuffing and padded claims, with PDF reports. Workspace, CLI and Windows app. Scores visible text only. Demo: vetta-vet-applicants.netlify.app | Proprietary, use requires permission: thisisanoopshekhar89@gmail.com

Official repository for StegPDF-21, providing the complete experimental pipeline for PDF steganography detection, including dataset generation, feature extraction, machine learning benchmarking, documentation, and links to the publicly available dataset.

Python toolkit building intelligent document containers — ordinary .docx and .odt files that also carry signed, encrypted payloads through five OOXML/ODF channels: custom XML parts, package parts, metadata properties, image LSB steganography, and MCE AlternateContent. Ed25519 manifests, AES-256-GCM, Merkle binding.

Reads a PDF and reports what is still inside it that you may not mean to send: text under a box that was never removed, an earlier version of the file, an attachment, a name in the properties. Runs entirely on your machine.

Prevent signature theft through cryptographic binding and steganography

See exactly what the model reads. Open-source document trust scanner and benchmark for AI pipelines.

Research code for robust local-region image watermarking with adaptive geometric correction.

Python tool for adding password protection to PDF files. Secure your sensitive documents with encryption.

A malicious forensic triage toolkit for examining Office and PDF files for advanced forms of malware payloads before execution in sensitive environments.

Finds prompt-injection payloads hidden inside resume PDFs — text a human reviewer never sees but an AI screener reads and acts on. Five independent detection layers.