Android clone and fake-app analyser
A static-analysis system for comparing an Android application package with a trusted reference and identifying signs of cloning, repackaging, or misleading identity.
Project definition
Problem statement
A fake or repackaged Android application may copy a legitimate name, icon, screens, or code while changing its signing certificate, permissions, embedded endpoints, or behaviour.
The engineering problem is to compare packages using multiple static signals and present traceable evidence without treating visual similarity or a single permission as proof of malicious intent.
Project objectives
- Extract package identity, signing certificates, manifest data, permissions, components, resources, and selected code features.
- Compare a candidate APK with a trusted reference package and version.
- Identify certificate mismatches, suspicious permission changes, copied resources, and changed network endpoints.
- Produce an explainable risk score supported by individual findings.
- Measure detection and false-positive performance on prepared benign and repackaged samples.
System design
System modules
Package intake
Validates APK files, calculates hashes, records provenance, and processes each sample in an isolated job.
Static extractor
Reads certificates, manifest entries, permissions, components, resources, strings, and selected code metadata.
Comparison engine
Compares package identity, certificate chains, icons, resources, permissions, components, and endpoints.
Risk scorer
Combines documented weighted signals and shows how every signal changes the result.
Evidence dashboard
Displays side-by-side differences, matched resources, warnings, confidence, and exported findings.
Methodology
System workflow
- 01Register reference
A trusted package is stored with its source, version, hashes, and expected certificate information.
- 02Upload candidate
A prepared candidate APK is hashed and sent to the isolated static-analysis worker.
- 03Extract features
The system records identity, certificate, manifest, permission, resource, and endpoint features.
- 04Compare
Differences and similarities are scored under a versioned rule set.
- 05Review evidence
The analyst inspects each signal and records a final classification for evaluation.
Demonstration scenario
A trusted sample and a prepared repackaged variant are compared. The analyser shows a different signing certificate, copied icon resources, two added permissions, and a changed endpoint. The risk score links to each difference while the reviewer records the final sample label.
Engineering
Technical architecture
- Web application
- Nuxt and Vue for sample intake, side-by-side comparison, finding review, and result export.
- Analysis API
- FastAPI for package jobs, extracted features, comparisons, risk results, and analyst labels.
- Data layer
- PostgreSQL for package metadata, hashes, certificates, features, comparisons, findings, and labels.
- Static analysis
- Python and Androguard for safe package parsing, with perceptual hashing for selected image-resource comparisons.
- Isolation
- Containerised workers, strict file limits, no application execution, and no outbound access during analysis.
Testing
Evaluation
Evaluation measures
- Accuracy on prepared trusted, benign variant, cloned, and repackaged samples
- Precision and recall for certificate, permission, resource, and endpoint findings
- False-positive rate across legitimate version changes
- Agreement between calculated risk and documented analyst labels
- Evidence traceability from score to extracted feature
- Analysis time and failure handling across package sizes
System boundaries
- Only applications owned by the student, safely prepared samples, or packages with permission for analysis are included.
- The system performs static analysis and does not execute unknown applications.
- A high score indicates evidence requiring review and is not a legal finding or malware verdict.
- The project does not collect applications from devices or app stores automatically.
Included
- 01APK upload, comparison, and evidence dashboard
- 02Certificate, manifest, resource, and code-feature analysis
- 03Explainable similarity and risk-scoring module
- 04Prepared safe APK samples, comparisons, and test results
- 05Complete source code in a private GitHub repository
- 06Complete project documentation with synopsis, abstract, methodology, analysis diagrams, test results, screenshots, and conclusion
- 07Setup and usage guide
Project record
No buyer information is collected on this page.
- Permanent project ID
- GP-CY-0ABQ3M8
- Catalogued
- 21 Aug 2026
- Completed
- Pending
- Verified
- Pending
- Demonstration
- Added when ready
Handover
After purchase
- 01Payment is confirmed
The project is marked unavailable and cannot be purchased again.
- 02Repository access is granted
The buyer's submitted GitHub account receives access to the private repository.
- 03The purchase record is delivered
The certification sheet is prepared from the reviewed buyer details and sent privately by email.