Secure development lifecycle
Security should be considered from requirements and design through implementation, testing, deployment and maintenance. The depth of review should be proportionate to the sensitivity of the data, exposure of the service and potential operational impact.
Design and change planning
- Identify trust boundaries, sensitive data, external integrations and administrative functions.
- Consider authentication, authorisation, tenant separation, input validation, logging and failure handling.
- Review material architecture or data-flow changes for security and privacy implications.
- Record security requirements and unresolved risks before release decisions.
Code and dependency controls
Proposed engineering practices include peer review for material changes, protected source repositories, restricted write permissions, secret scanning, dependency review and timely updates for vulnerable packages. Python projects may use tools such as Bandit and pip-audit; container images may be assessed with Trivy; secret exposure may be checked with Gitleaks. Tool use and scan results must be evidenced before claiming that a particular repository or release is covered.
Testing and release
Testing should include relevant unit and integration tests, authentication and authorisation checks, input validation, error handling and regression tests. Security testing should be performed in authorised non-production environments where possible. Releases should have an identifiable change, review and rollback approach appropriate to the service.
Configuration and secrets
Environment-specific configuration should be separated from application code. Secrets should be stored in an approved secret-management mechanism or protected deployment configuration, access-restricted and rotated when exposure is suspected. Debug interfaces and development credentials should not be exposed in public deployments.
Maintenance
After release, owners should monitor relevant security advisories, assess affected dependencies and base images, apply fixes according to risk, and retain evidence of remediation and verification.
Evidence and limitations
Potential evidence includes repository protection settings, pull-request reviews, CI results, dependency and container scan reports, release records and remediation tickets. The current toolchain, coverage, enforcement and scan cadence across Apps Intellect products remain subject to source-code and deployment review.
