CogniPass v1.0

CogniPass is a high-speed, targeted password generator and OSINT profiler written in Zig. It does not just blindly brute-force characters; instead, it uses cognitive chunking theory to generate passwords the way humans naturally invent them—by mangling and combining Personal Identifiable Information (PII) like names, birthdates, spouses, kids, and pets.

Features

• Interactive Profiling: Automatically prompts for target PII (names, dates, pets) and handles validation (rejects Cyrillic and special characters).
• CLI Mode: Fully automatable via command-line arguments.
• Cognitive Levels: Generates passwords grouped by how the human brain creates them:
  • Level 0: Global Baseline (e.g., 123456, password)
  • Level 1: High Availability / Chunking (Raw PII)
  • Level 2: Low Cognitive Load (Simple Caps & Dates)
  • Level 3: Illusion of Complexity (Special chars & separators)
  • Level 4: Security Fatigue (Token Mangling & L33t Speak)
• Target Combination Count: Will smartly mutate and combine tokens to meet a specific dictionary size if requested.
• File Overwrite Protection: Built-in safety checks for generating dictionaries.

---

🚀 Building for Maximum Performance

Zig is chosen for its extreme execution speed and lack of hidden control flow. However, maximum speed comes from a combination of compiler flags and code architecture.

(Note: Requires Zig 0.13.0+)

1. The Ultimate Fast Build (Local Machine)

For the absolute fastest execution on your specific CPU architecture (enables AVX2, BMI2, etc. if available):

zig build-exe src/main.zig -O ReleaseFast -mcpu=native

Warning: The resulting binary might not run on older CPUs due to -mcpu=native.

2. Portable Fast Release (For Distribution)

If you need to distribute the binary to other machines:

zig build-exe src/main.zig -O ReleaseFast

3. Safe Build (For Development & Testing)

Leaves runtime safety checks enabled (array bounds, unhandled errors). Useful when modifying the rule engine:

zig build-exe src/main.zig -O ReleaseSafe -mcpu=native

---

🧠 Prediction Logic & Feature Roadmap

Now that the core engine is fully optimized (Zero-Allocation, Batch I/O, Memcpy concatenations), future updates will focus on improving the prediction logic and functionality:

1. Expanded Family Tree & Relational Graph

• Add logic for parent/sibling dates.
• Cross-relational combinations (e.g., SpouseName + ChildBirthYear).

2. Intelligent Keyword Mangling

• Context-aware L33t Speak (e.g., replace 'i' with '1' only in the middle of a word).
• Transpositions and common typo variations (e.g., michael -> micheal).
• Keyboard walk patterns combined with PII (e.g., qwer + dob).

3. Localization & Culture-Specific Rules

• Date format awareness (DD.MM.YYYY vs MM/DD/YYYY) based on locale.
• Localized keyboard layouts (e.g., QWERTY vs AZERTY transpositions).

4. Advanced Output Formatting

• Hashcat/John the Ripper specific formatting or direct integration.
• Entropy scoring for each generated candidate (sorting the dictionary by probability).

---

Usage Examples

Interactive Mode:

./Passgen.exe

CLI Mode:

./Passgen.exe --first michael --last garcia --dob 12.07.1984 --spouse-first jessica --spouse-dob 05.09.1986 --min-len 8 -c 1000 --out custom_dict.txt

PIN Only Mode:

./Passgen.exe --dob 12.07.1984 --spouse-dob 05.09.1986 --pin --out pins.txt