Kembali ke Blog
Info 10 menit

Security Audit Akunpy — 2 Tahun Operasi, Zero Data Breach (Mei 2026)

Per Mei 2026, Akunpy operate 2 tahun (sejak Maret 2024) dengan track record: zero data breach, zero credential leak, zero compromise critical. Halaman ini detail security infrastructure untuk transparency publik + audit reference.

Note: dokumen ini designed untuk security researcher, enterprise customer compliance audit, akademisi yang research sharing economy security. Public CC-BY-4.0 license.

Security Architecture Overview

Akunpy security architecture multi-layer defense:

1. Transport Layer: TLS 1.3 + HSTS preload

2. Authentication: JWT RS256 + rotation + bcrypt password hashing

3. Data Layer: AES-256-GCM encryption at rest

4. Slot Isolation: Per-user encryption keys + password lock

5. Anti-Abuse: Rate limiting + anomaly detection

6. Monitoring: 24/7 SOC + automated alerting

7. Incident Response: Documented runbook + practice quarterly

8. Audit: Quarterly internal + annual external

Layer 1: Transport Security (TLS 1.3)

Spec:

Protocol: TLS 1.3 ONLY (TLS 1.0/1.1/1.2 disabled)

Cipher: AES-256-GCM with ECDHE key exchange

Certificate: Cloudflare Universal SSL + Let's Encrypt fallback

HSTS: max-age=63072000 (2 tahun) + includeSubDomains + preload

HSTS Preload List: ✓ (registered Sept 2024)

Test result (SSL Labs A+): https://www.ssllabs.com/ssltest/analyze.html?d=akunpy.com

Layer 2: Authentication (JWT + bcrypt)

Spec:

JWT signing: RS256 (asymmetric, public/private key)

Token rotation: every 24 jam aktif

Max session: 7 hari inaktif → auto-logout

Password hashing: bcrypt cost 12 (current standard 2026)

Salt: per-password random salt

Optional MFA: TOTP support (Google Authenticator, Authy)

Test: try inspect cookies di browser. HttpOnly + Secure + SameSite=Strict flags semua set.

Layer 3: Data Encryption (AES-256-GCM)

Spec:

Algorithm: AES-256-GCM (Authenticated Encryption with Associated Data)

Key derivation: PBKDF2 dengan 600.000 iterations

Key rotation: every 90 hari

Encryption keys: stored di Cloudflare D1 encrypted + master key di HSM

Backup encryption: same key hierarchy, backup ke S3 dengan KMS

Coverage:

All user credentials

Slot assignment data

Session tokens

Payment metadata (post-Midtrans/Xendit processing)

Audit log entries

Layer 4: Slot Isolation (Akunpy Innovation 2024)

Per-user slot isolation system:

Setiap user dapat slot eksklusif (private encryption key)

Password lock individual (mandatory untuk Claude access)

Chat history not shared di Akunpy server (proxy only)

Cross-slot access impossible (cryptographically isolated)

Validation:

Penetration test Q4 2024: pass

Penetration test Q4 2025: pass

External audit Q2 2026: pass

Real-world: zero cross-user data leak in 2 tahun

Layer 5: Anti-Abuse System

Rate limiting:

Per user: 13 messages / 5 hour rolling (sesuai claude.ai standard 2026)

Per IP: max 100 req/second (DDoS protection)

Per device: fingerprint-based untuk prevent multi-account abuse

Anomaly detection:

Login pattern analysis (geographic, time-based)

Unusual usage spike (>5x normal pattern)

Fraud signal (mismatched user agent, suspicious payment)

Auto-suspend trigger threshold tertentu

False positive rate: <0.5% (legitimate users accidentally flagged) — appeal via support unlock dalam 5 menit.

Layer 6: Monitoring & Alerting

Real-time monitoring:

Grafana dashboard 24/7 untuk system metrics

Telegram alert untuk: error spike, unusual login pattern, payment failure burst

PagerDuty-equivalent on-call rotation untuk critical incident

Log aggregation: 30-day hot storage + 1-year cold storage

Coverage:

All authentication events

All payment events

All admin actions (audit trail)

All API errors

Alert threshold:

Critical: page within 5 menit (e.g., database down, payment gateway failure)

Warning: notify within 15 menit (e.g., slow query, unusual error rate)

Info: digest daily (e.g., new user signup volume)

Layer 7: Incident Response

Documented runbook untuk berbagai scenario:

1. Data breach (suspected/confirmed)

2. Account compromise (mass)

3. Payment fraud incident

4. DDoS attack

5. Database corruption

6. Critical bug in production

Process:

1. Detect (monitoring/alert/user report)

2. Confirm (within 15 menit)

3. Contain (within 1 jam — isolate affected components)

4. Investigate (within 4 jam — root cause)

5. Remediate (within 24 jam — fix + deploy)

6. Communicate (within 48 jam — post-mortem public kalau impact user)

7. Learn (post-mortem analysis + improvement)

Practice: tabletop exercise quarterly. Real incident drill annual.

Track Record 2 Tahun (Maret 2024 - Mei 2026)

Incident type breakdown:

Critical incident (data breach, full outage): 0

Major incident (partial outage >1 hour): 3 (Q4 2024 deploy issue, Q2 2025 DDoS, Q4 2025 Cloudflare network)

Minor incident (degraded performance): ~30 (sebagian besar resolve <30 menit)

Customer-impacting bugs: ~50 (avg time-to-fix 4 jam)

Outage time total tercatat rendah selama 2 tahun operasi — gangguan slot dipulihkan otomatis via auto-recovery.

Customer compensation (extend subscription) saat outage major: 3 incidents = total ~6 hari extend untuk 8K+ users impacted.

Audit History

Q4 2024: External security audit pertama. Vendor: Indonesian security firm. Scope: infrastructure + application code review. Findings: 0 critical, 2 medium (resolved within 30 hari), 5 low (resolved within 90 hari). Report: available via NDA untuk enterprise customer.

Q4 2025: External security audit kedua. Vendor: Singapore-based regional firm. Scope: expanded include penetration test + social engineering. Findings: 0 critical, 1 medium (resolved within 14 hari), 3 low (resolved within 60 hari).

Q2 2026 (Mei): Internal comprehensive audit. Findings: 0 critical, 0 medium, 4 low (improvement opportunities). Report: this document.

Future audit: SOC 2 Type II audit planned Q4 2026 untuk enterprise customer requirement.

Compliance Status

Akunpy compliant dengan:

✓ UU ITE 2008/2016 (Indonesia)

✓ UU Perlindungan Konsumen 1999

✓ UU PDP 2022 (Personal Data Protection)

✓ PP 71/2019 (PSE registration)

✓ OWASP Top 10 2021 mitigations

In progress:

⏳ SOC 2 Type II certification (target Q4 2026)

⏳ ISO 27001 certification (target 2027)

⏳ Pre-audit untuk EU GDPR compliance (untuk global expansion 2027+)

Bug Bounty Program

Akunpy welcome responsible disclosure dari security researchers. Reach: support@akunpy.com dengan subject "Security Disclosure".

Reward (per severity):

Critical (data breach, RCE, auth bypass): Rp 5-15 juta

High (privilege escalation, sensitive info disclosure): Rp 1-5 juta

Medium (CSRF, XSS, SQLi mitigated): Rp 500K-1 juta

Low (info leak, version disclosure): Rp 100-500K

Honorable mention (out-of-scope or low-impact): Akunpy Pro 1 tahun free + public credit

Hall of Fame: /keamanan/hall-of-fame (kalau ada — currently empty karena 0 valid bug bounty in 2 years).

Compare to Industry Average

Sharing platform industry (informal Discord/Telegram resellers):

Data breach: multiple reported incidents per year per platform

Encryption: often weak or plaintext

Audit: typically none

Monitoring: ad-hoc

Incident response: undefined

Enterprise SaaS standard (Anthropic, OpenAI):

Data breach: occasional (e.g., OpenAI March 2023 ChatGPT exposed user data 9 jam)

Encryption: AES-256+

Audit: SOC 2 Type II + ISO 27001

Monitoring: 24/7 SOC professional

Incident response: well-defined

Akunpy positioning: aspire to enterprise SaaS standard. Already above sharing industry standard. Gap closing dengan certification target 2026-2027.

Untuk Enterprise Customer

Compliance documentation available under NDA:

Detailed security architecture diagram

Latest external audit report

Data Processing Agreement (DPA) template

Sub-processor list

Incident response runbook

Penetration test results

Hubungi partnership@akunpy.com untuk request.

Kesimpulan

Akunpy 2 tahun operasi dengan zero data breach + strong security architecture. Multi-layer defense, regular audit, transparent track record. Above-industry standard untuk sharing platform, approaching enterprise SaaS standard. Untuk enterprise customer, full compliance documentation tersedia under NDA. Trust verified by track record, bukan marketing claim.

Ada pertanyaan lain?

Buka Dashboard