iso27001-sgsi

ISO 27001/SGSI — CIA triad, controls, risk assessment.

SKILL.md

GitHub

ISO 27001 / SGSI / CIA Triad


CIA Triad

The three foundational properties of information security. Every security decision must be evaluated against all three.

Pillar Definition Software implications
Confidentiality Information is accessible only to authorized parties Encryption at rest and in transit, access control, least privilege, no PII in logs
Integrity Information is accurate and has not been tampered with Input validation, checksums, audit logs, signed tokens, immutable audit trails
Availability Systems and data are accessible when needed Rate limiting, DDoS mitigation, backups, graceful degradation, health checks, SLAs

When a control improves one pillar, check that it does not degrade another (e.g., aggressive rate limiting improves availability against DDoS but may harm legitimate users — balance required).


SGSI — Information Security Management System

An SGSI (Sistema de Gestão de Segurança da Informação / ISMS) is the framework of policies, processes, and controls that an organization uses to manage information security risk systematically.

Core cycle (Plan–Do–Check–Act):

Phase What happens
Plan Define security scope, conduct risk assessment, set objectives and controls
Do Implement selected controls (technical, operational, organizational)
Check Monitor, audit, measure effectiveness of controls
Act Review findings, apply corrective actions, improve continuously

For software teams, this translates to:

  • Threat modeling before building features that handle sensitive data
  • Risk register maintained and reviewed each sprint or quarter
  • Security review gate before production deployments (enforce via CI/CD)
  • Incident response plan documented and tested
  • Security training for all engineers handling customer data

ISO 27001 — Software-Relevant Controls (Annex A)

Focus on the controls most directly actionable by a development team:

A.8 — Asset Management

  • Classify data by sensitivity (public / internal / confidential / restricted)
  • Document which services process which data classifications
  • Define retention and disposal policy for each classification

A.9 — Access Control

  • Principle of least privilege: roles grant minimum permissions to accomplish the task
  • Privileged access (admin, prod DB) requires MFA + audit logging
  • Access rights reviewed quarterly; revoked immediately on offboarding
  • No shared credentials — every system actor has a unique identity

A.10 — Cryptography

  • Encryption in transit: TLS 1.2 minimum; TLS 1.3 preferred
  • Encryption at rest: AES-256 for confidential and restricted data
  • Key management: keys stored in a secret manager (AWS KMS, HashiCorp Vault, GCP KMS); never hardcoded
  • Password hashing: bcrypt (cost ≥ 12), Argon2id, or scrypt — never MD5, SHA-1, or unsalted SHA-2

A.12 — Operations Security

  • Change management: all production changes go through a review process (PRs, approvals)
  • Malware protection: dependency scanning, container image scanning in CI
  • Logging and monitoring: audit logs for authentication, authorization decisions, data access, and admin actions
  • Log retention: minimum 90 days hot, 1 year cold (adjust to regulatory requirements)

A.13 — Communications Security

  • Network segmentation: databases not directly reachable from the internet
  • API security: authentication required on all endpoints; no unauthenticated data exposure
  • Data transfer: PII/sensitive data transferred only over encrypted channels; no email attachments for sensitive data

A.14 — System Acquisition, Development, and Maintenance

  • Security requirements defined before development starts (threat model)
  • Secure coding standards enforced (OWASP Top 10, this skill set)
  • Security testing in CI: SAST, dependency audit, secret scanning
  • Separation of environments: dev/staging/production strictly isolated; no production data in dev/staging

A.16 — Incident Management

  • Incident response plan: defined roles, communication channels, escalation path
  • Security incidents logged and post-mortemed
  • Breach notification: know your legal obligation (LGPD: 72h to ANPD; GDPR: 72h to DPA)

A.18 — Compliance

  • Legal requirements identified: LGPD, GDPR, PCI-DSS, HIPAA — whichever apply to the project
  • Privacy by design: collect minimum data, document purpose, implement erasure
  • Regular compliance audits against applicable regulations

Risk Assessment Checklist (per feature / per sprint)

Ask these before building any feature that handles user data or system access:

  • What data does this feature collect or process? What is its classification?
  • Who can access this data? Is access scoped to the minimum necessary?
  • What happens if this data is exposed? (Confidentiality impact)
  • What happens if this data is corrupted or tampered with? (Integrity impact)
  • What happens if this feature is unavailable? (Availability impact)
  • Are there regulatory obligations attached to this data?
  • Are audit logs generated for sensitive actions in this feature?

Audit Log Requirements

Every audit log entry must capture:

Field Description
timestamp ISO 8601, UTC
actor_id Authenticated user/service identity
action What was done (order.created, user.login, admin.permission_granted)
resource_type Object type affected
resource_id Object ID affected
outcome success or failure
ip_address Client IP (where applicable)
user_agent Client context (where applicable)

Audit logs must be:

  • Append-only (no update or delete)
  • Stored separately from application logs
  • Tamper-evident (hash chaining or write to immutable storage)
  • Retained per regulatory requirement (minimum 1 year)