SERVICE #03 • COERCION RESISTANCE

Duress Decoy Vaults & Plausible Deniability

⏱️ 7 min read 🛡️ Dual AES-256-GCM Envelopes 📐 IND-CPA Indistinguishability

In hostile physical situations (such as border crossings, device confiscation, or physical coercion), standard cryptographic defenses fail because the adversary forces you to reveal your password. ZevSafe's Duress Decoy Vault provides plausible deniability: entering your secondary decoy password unlocks a convincing set of innocuous documents while keeping your true sensitive data mathematically invisible.

🟢 Simple User Guide

How Duress Mode Protects You

STEP 01

Enable Duress Mode During Lock

When creating your vault, toggle "Duress Decoy Vault". Enter your confidential Primary Password, and then define a completely different Decoy Password.

STEP 02

Select or Generate Decoy Files

Provide innocent files (such as family vacation photos, tax summaries, or receipts) or click "Generate Realistic Decoys" to create convincing travel itineraries and household budgets automatically.

STEP 03

Unlock According to Context

To access your true files, enter your Primary Password. If forced to surrender your password under physical coercion, enter your Decoy Password. ZevSafe unlocks the decoy vault cleanly—leaving zero indication that another vault exists!

⚠️
Operational Security (OpSec) Tip: Make sure your decoy password is easy to remember and that your decoy files match your real-world identity (e.g. realistic dates, typical household documents).
🔵 Technical Specification

Dual-Manifest Container Architecture

When Duress Mode is active, ZevSafe sets bit 1 in the header flags (V3_FLAG_DURESS = 0x02) and appends two independent encrypted manifests at the tail of the container:

Duress Trailer Layout 
+-------------------------------------------------------------------------+
| Primary Data Chunks + Decoy Data Chunks (Interleaved STREAM AEAD)       |
+-------------------------------------------------------------------------+
| ENVELOPE 0 (Primary Manifest):                                          |
| [ Salt (32B) | IV (12B) | Length (4B) | Ciphertext (M0) | GCM-Tag (16B) ]  |
+-------------------------------------------------------------------------+
| ENVELOPE 1 (Decoy Manifest):                                            |
| [ Salt (32B) | IV (12B) | Length (4B) | Ciphertext (M1) | GCM-Tag (16B) ]  |
+-------------------------------------------------------------------------+

Each envelope possesses its own independent 32-byte CSPRNG salt, 12-byte initialization vector, and 128-bit authentication tag.

Mathematical Indistinguishability ($IND\text{-}CPA$)

Under the standard cryptographic security definitions of AES-256-GCM, the ciphertext output is computationally indistinguishable from uniform random noise:

📐
$$\mathbf{Adv}_{\text{AES-GCM}}^{\text{ind-cpa}}(\mathcal{A}) \le \epsilon \approx 2^{-128}$$
When an examiner supplies the decoy password, Envelope 1 validates its authentication tag and decrypts the decoy catalog. Envelope 0, meanwhile, fails tag validation and is indistinguishable from pseudorandom padding. The examiner has no mathematical mechanism to determine whether Envelope 0 represents an unauthenticated primary vault or simply random noise.
Dual Envelope Authentication Routing 
// js/stream-unpacker.js: Authenticate candidate password against both envelopes
async function unlockDuressVault(trailerBytes, candidatePassword) {
    // 1. Attempt decryption of Envelope 0 (Primary)
    try {
        const manifest = await decryptEnvelope(envelope0Bytes, candidatePassword);
        return { type: 'PRIMARY', manifest };
    } catch (e) {
        // Authentication tag failed — expected if candidate is decoy password
    }

    // 2. Attempt decryption of Envelope 1 (Decoy)
    try {
        const manifest = await decryptEnvelope(envelope1Bytes, candidatePassword);
        return { type: 'DECOY', manifest };
    } catch (e) {
        throw new Error("Invalid password: Failed authentication across both envelopes");
    }
}

Automated Realistic Decoy Generator

Empty vaults or single dummy files trigger suspicion during forensic inspections. ZevSafe includes an automated synthetic generator that constructs realistic file structures:

Category Generated Synthetic Files Realistic Content
Travel & Vacations itinerary.txt, hotel_confirmation.pdf Flights, booking references, rental car details, day-by-day sightseeing.
Culinary & Recipes sourdough_guide.md, mediterranean_diet.txt Ingredient lists, fermentation temperatures, baker's percentages.
Personal Budget monthly_expenses_2026.csv, utilities.txt Plausible electricity, water, rent, and grocery line items.

Security Invariants

Invariant Guarantee Audit Result
Cross-Envelope Isolation Decoy key cannot decrypt primary data chunks or manifest. VERIFIED • Distinct PBKDF2 salts and keys.
Zero UI Tell-Tale Signs Unlocking in decoy mode presents identical explorer UI without warning badges. VERIFIED • Indistinguishable user interface.
Safe Buffer Scrubbing Candidate key material zeroized after failed attempts. VERIFIED • .fill(0) on all intermediate buffers.