SkillGap AI — Career Intelligence Platform

Analyze your resume against any job description and get an instant, AI-powered skills-gap report: employability score, matched & missing skills, role compatibility across 9 career blueprints, a personalized 12-week learning roadmap, and an AI mentor you can chat with.

What the app does

Feature	Description
Auth	Sign-in / register / reset-password with PBKDF2-SHA256 hashing, 5-attempt lockout, password-strength meter, seeded demo accounts
Skill Analyzer	Upload a resume (PDF/TXT) or paste text, pick a job role template → extracted skills vs. ~120-skill taxonomy → employability gauge, match %, matched/missing/extra skill clouds, resume suggestions
Role Analytics	Compatibility score + salary estimate for Full-Stack, Frontend, Backend, ML, Data, Cloud/DevOps, Cyber Security, Mobile, AI/LLM roles
Roadmap	Auto-generated phased learning plan with curated resources and progress tracking
AI Mentor	Chat assistant — Gemini 2.0 Flash when an API key is configured, offline expert fallback otherwise
Settings	Profile, target role, engine status

Repository layout

├── app.py                  # Streamlit app (primary web UI, port 8501)
├── backend.py              # Core logic: auth, resume analysis, mentor
├── database.py             # SQLite user store (skillgap_users.db)
├── index.html + app.js + style.css   # Static vanilla-JS frontend
├── api/                    # Vercel serverless functions (analyze, chat)
├── skillgap_flutter/       # Flutter mobile/web/desktop port (Dart)
├── selenium-tests/         # Mocha + WebDriver E2E suite (~300 cases)
├── appium-tests/           # WebdriverIO + UiAutomator2 mobile suite
├── load-tests/             # k6 load test (100 VUs, GET-only)
├── scripts/                # Excel report generators
├── Vulnerability Test Results/   # Security scan outputs
└── .github/workflows/      # CI: test_suite.yml, all-tests.yml, build-and-deploy.yml, security-review.yml

---

A–Z: run everything

0. Prerequisites

Tool	Version	Install
Python	3.11+	python.org
Node.js	20+	nodejs.org
Flutter	stable	flutter doctor
Chrome	latest	for Selenium
k6	latest	choco install k6 / grafana.com/docs/k6
Appium 2	optional	npm i -g appium + appium driver install uiautomator2

1. Python dependencies (once)

pip install -r requirements.txt

(streamlit pypdf openai openpyxl bcrypt python-dotenv)

2. Run the Streamlit web app

streamlit run app.py

Opens at http://localhost:8501. This is the main UI and the target of the Selenium / load-test suites.

3. Run the static frontend

python -m http.server 8080

then open http://localhost:8080/index.html. The api/ serverless functions run on Vercel (vercel dev), not in this static server.

4. Demo accounts

Email / User	Password	Profile
[email protected]	Password123!	Full demo analysis
alex	Password123!	Mid-level developer
priya	Password123!	Data-science track

5. Run the Flutter app

cd skillgap_flutter
flutter pub get

flutter run -d chrome      # web
flutter run -d windows     # windows desktop
flutter run                # connected Android device/emulator

Build artifacts:

flutter build apk --release     # → build/app/outputs/flutter-apk/app-release.apk
flutter build web --release     # → build/web/
flutter analyze && flutter test # lint + unit tests

6. Run the Selenium suite (needs app on :8501 + Chrome)

streamlit run app.py &            # terminal 1
cd selenium-tests
npm install
npm test                          # terminal 2 — mocha spec output
npm run test:full                 # JSON results + Excel report

Reports land in selenium-tests/reports/ + selenium-test-summary.xlsx.

7. Run the Appium suite (needs Android emulator/device)

appium --port 4723 &              # terminal 1
cd appium-tests
npm install
BASE_URL=http://localhost:8501 npm test
npm run report                    # → appium-test-report.xlsx

Default target: Android 13 emulator emulator-5554 against https://skill-gap-ai.streamlit.app — override with BASE_URL.

8. Run the k6 load test

streamlit run app.py &            # app must be reachable
k6 run --env BASE_URL=http://localhost:8501 load-tests/load-test.js

Thresholds: p50 < 500 ms, p90 < 1500 ms, p95 < 2500 ms, p99 < 5000 ms, error & HTTP-failure rates < 5%. Reports → load-tests/reports/.

9. Generate Excel reports

python scripts/generate_excel_reports.py   # findings + endpoint inventory
python scripts/generate_test_excels.py     # selenium + appium workbooks

10. Security scans

pip install bandit pip-audit semgrep
bandit -r .
pip-audit -r requirements.txt
semgrep scan --config auto

Outputs are collected under Vulnerability Test Results/.

---

CI / GitHub Actions

Pushing to main runs test_suite.yml:

setup-node ─┬─ Unit Tests — API (300) ─────→ Deployment Status (300) ─┐
            ├─ Selenium — Website Tests (300)                          │
            ├─ Appium — Android Tests (300)                            ├─→ Compile Master Report & Deploy
            ├─ Validation Tests (300)                                  │
            ├─ Load Testing — Performance (300)                        │
            ├─ security-assessment                                     │
            └─ web-build (Flutter web)                                 ┘

Artifacts per run: web-build, selenium-report, appium-report, load-report, security-report, unit-test-report, validation-report, deployment-status-report, master-report (90-day retention).

Other workflows: all-tests.yml (security audit + all 4 suites + Excel), build-and-deploy.yml (APK + website deploy), security-review.yml.