Android Reset Lab

A simulation-only security lab for authorizing, approving, recording, and verifying high-impact device reset and sanitization workflows.

No real device commands are executed. The lab exists to test controls around irreversible operations without putting user data or evidence at risk.

What it demonstrates

• Default-deny RBAC and time-limited sessions
• Distinct-requester dual control for reset approval
• Risk-adaptive step-up and transaction signing
• Device attestation and policy-as-code decisions
• Hash-chained, HMAC-protected and Merkle-verifiable audit evidence
• Detection for brute force, replay, privilege abuse and log tampering
• Safety classification across troubleshooting, redeployment, loss, investigation and retirement
• Clear separation between factory reset, cryptographic purge, evidence preservation and physical destruction

Safety model

The sanitization advisor evaluates:

• operational intent
• evidence or legal hold
• encryption from first use
• removable media
• Factory Reset Protection or activation lock
• device connectivity
• required assurance: Clear, Purge or Destroy

It returns one of five deterministic outcomes:

BLOCK · PRESERVE · SIMULATE_RESET · SANITIZE · DESTROY

A factory reset is never presented as universal proof of sanitization. Investigation scenarios preserve evidence, removable media is handled separately, and Purge requires a defensible encryption or approved sanitization path.

Run locally

python -m pip install -r requirements.txt
python seed_lab.py
python web_console.py

Open http://localhost:8080.

Seeded simulation accounts are defined in seed_lab.py. Change all credentials before using the lab outside an isolated local demonstration.

Verify

python -m pytest -q
python -m compileall -q . -x '(.git|.venv|venv|__pycache__)'
python -m pip check

The safety test statically rejects destructive Python calls and keeps SIMULATION_MODE mandatory.

Architecture

Authentication → authorization → risk/policy decision
       → reset request → distinct approval → simulated execution
       → signed audit event → detection → verification/reporting

Important modules:

• cerberus_workflow.py — orchestrated request/approval/execution controls
• sanitization_advisor.py — deterministic intent and assurance guidance
• policy_engine.py — default-deny policy decisions
• risk_engine.py — contextual risk scoring
• merkle_ledger.py — transparency and inclusion evidence
• security_logger.py — tamper-evident security events
• web_console.py — simulation console

Scope boundary

This project does not connect to ADB, Android Management API, carrier services, or real devices. A production system would additionally require durable storage, managed keys, independent identities, external approval delivery, device-vendor validation, legal retention controls, and verified media-sanitization procedures.

Security

Report vulnerabilities privately according to SECURITY.md. Do not include live credentials, personal information, or production device identifiers.

License

See LICENSE.