Security as a
written contract

Security is not a posture; it's a set of written controls, mapped to evidence, tested on a schedule, reviewed after every incident. We build that layer into the product, not around it.

Threat model: internet edge, DMZ, trust zones, attack paths, detection panel threat.model edge · dmz · trust attack paths · detection Internet · edge DMZ App tier WAF API gateway Auth App · svc DB · encrypted attacker Detection · last 24h SIEM rules fired 148 High-severity paged 2 False-positive rate 3.1% MTTD · median 4 min

CAPABILITIES

Six surfaces under one security contract

Audit, access, vulnerabilities, cloud risk, detection, compliance — engineered together. Every control traces to written policy and measurable evidence.

01

Audit & risk mapping

Threat model, asset inventory, trust boundary diagrams, likelihood × impact scoring. The written basis for every security decision that follows.

02

Access policies · IAM

Least-privilege across clouds, SSO with step-up auth, SCIM, break-glass procedures. Who-can-do-what written in code and tested in CI.

03

Vulnerability review

SAST, DAST, SCA, IaC scanning, supply-chain integrity. Findings triaged by blast radius, not by CVSS alone.

04

Cloud risk

Guardrails across AWS / GCP / Azure, encryption defaults, network segmentation, managed-service patterns reviewed per workload.

05

Detection & response

Logs, EDR, SIEM rules that actually fire on the right signals. Pages that wake the right person with context, not noise.

06

Compliance mapping

SOC 2, ISO 27001, HIPAA, PCI-DSS, GDPR, EU AI Act. Controls traced to evidence — ready for the next audit without a sprint of screenshot archaeology.

FINDINGS LEDGER

Six findings from a real audit, ranked and actioned

A worked example of the findings ledger we hand back at the end of an audit. Every row names an action — not a generic "review policy". Severity is blast-radius ranked, not CVSS alone.

Critical

Public S3 bucket with prod credentials

Cloud · IAM

Bucket policy · rotate credentials · audit access log

High

SAML group mapping bypass on SaaS-X

Access · IAM

SCIM sync · step-up auth · break-glass drill

High

Unpatched CVE in transitive dependency

Supply chain

Renovate · SCA · pinned lockfiles · release gate

Medium

Secrets in CI logs under sparse redaction

CI · secrets

Centralised secret store · log redaction · policy

Medium

Overly permissive IAM role on workloads

Cloud · IAM

Policy tightening · access analyser · least privilege

Low

Verbose error pages reveal stack frames

App · errors

Error sanitisation · observability → errors to backend only

Controls in code

Access policy, network segmentation, data protection and supply-chain controls work only if they live in code and tests, not in a PDF.

Open DevOps & cloud ↗

CONTROL LANES

Identity · network · data · supply chain

The four lanes every hardening engagement runs on. Each item is a shipped control — we don't recommend what we don't build.

Identity

  • SSO · SCIM · MFA / WebAuthn
  • Step-up auth for sensitive ops
  • Break-glass rotation drill
  • Audit every admin action

Network

  • VPC segmentation · private subnets
  • Zero-trust access (Cloudflare, Tailscale)
  • WAF · rate limits · bot detection
  • Egress controls + allowlists

Data

  • At-rest + in-transit encryption
  • KMS · envelope encryption
  • PII minimisation · tokenisation
  • Retention + delete-on-request

Supply chain

  • Signed artefacts · SBOM · SLSA
  • Dependency pinning · Renovate
  • GitHub advanced security
  • SCA across all registries

Adjacent disciplines

Audit · harden · detect

Put security in writing, in code and in the release pipeline

Share the architecture, the compliance profile and the incident history (under NDA). We come back with a threat model, findings ledger and remediation calendar inside ten working days.