ZeeshanSaud/CodeTruthAgent-V3-Module1-Evaluation

1---2license: gpl-3.03tags:4  - code-analysis5  - repository-classification6  - software-engineering7  - static-analysis8---9## Source Code10https://github.com/Zeeshan78699/CodeTruthAgent — tag `v3.0.0-module1`11 12# CodeTruth Agent V3 — Module 1 Evaluation13 14Validation results for **Module 1: Repository Cognition Engine** —15a deterministic, rule-based engine that scans a software repository16and determines its application type, primary framework, technology17stack, and file inventory.18 19## What's in this dataset20 21- `FULL_DOMAIN_SUMMARY.md` — summary table of all 69 validated repositories22- `FULL_DOMAIN_SUMMARY.json` / `.csv` — same data in structured form23- 69 individual per-repository scan reports (`.md`)24 25## Validation Results26 27```2869 repositories scanned2969/69 = 100% discovery score3069/69 = correct application type3169/69 = correct primary framework (or correctly "No Framework Detected")3269/69 = governance gate APPROVED33 0/69 = crashes34 0/69 = skipped3557/69 = 100% classification score3612/69 = 75% classification score ("No Framework Detected" — correct37        by design for non-Python system software)3839 distinct application types covered39441,660 total files scanned4035/35 unit tests pass41```42 43Repository sizes ranged from 35 files (python-sgp4) to 61,850 files44(Zephyr RTOS), including non-Python repositories (Redis, Nginx, Go,45Rust, U-Boot).46 47## Notes48 49"No Framework Detected" is a genuine, correct result for repositories50with no Python package framework dependency — it is not an error or a51gap.52 53## Source Code54 55github.com/Zeeshan78699/CodeTruthAgent — tag `v3.0.0-module1`56 57## Author58 59Zeeshan Saud — Independent AI Researcher, UAE