mitre-defend-review

Review a repository, architecture, workflow, or local folder with a focus on MITRE D3FEND and defensive coverage concepts. Use when Codex needs to assess how well a system detects, prevents, contains,

SKILL.md

GitHub

MITRE D3FEND Review

Overview

Use this skill to analyze a system through the lens of defensive coverage. Focus on which protections, hardening choices, visibility points, containment measures, and response-enabling controls exist by design, and which attacker behaviors appear weakly defended.

Prefer practical defensive coverage analysis over a long catalog of possible controls.

Workflow

1. Establish the review scope

Choose the smallest meaningful slice before reading everything.

Start with:

  • the README and architecture docs
  • dependency manifests and framework config
  • obvious trust boundaries, control points, observability paths, and privileged workflows

If the folder or codebase is large, choose one of these scopes and say which one you chose:

  • one exposed service or ingress path
  • one privileged workflow
  • one identity or secret-handling path
  • one end-to-end detection and containment path

2. Identify control surfaces

Identify where the system appears to:

  • authenticate and authorize
  • validate, sanitize, or constrain input
  • limit privilege or isolate tenants/components
  • protect secrets, tokens, and keys
  • log, trace, alert, or audit sensitive behavior
  • rate limit, retry safely, contain failures, or degrade safely
  • verify integrity of code, config, artifacts, or data

3. Map defensive coverage

Review the system for defensive coverage themes such as:

  • hardening and safe defaults
  • identity assurance and least privilege
  • integrity protection
  • network and transport protection
  • runtime containment and segmentation
  • auditability and monitoring
  • resilience, recovery, and operational containment

When helpful, describe controls in D3FEND-style language, but keep the output readable and practical.

4. Review the key design points

For each important control point, capture:

  • file path or architectural component
  • defensive purpose
  • what attacker behavior it may prevent, detect, or contain
  • what assumptions it relies on
  • what gaps remain if the control fails or is bypassed

Prioritize:

  • authentication and authorization gates
  • secrets and trust material handling
  • logging and audit paths
  • configuration defaults and fallback behavior
  • workflow approvals and privileged automation
  • segmentation, isolation, and blast-radius controls

5. Surface defensive gaps

Call out issues that materially weaken defensive posture, especially:

  • missing or optional controls on critical paths
  • weak detection for sensitive or attacker-relevant actions
  • poor isolation, broad privilege, or weak blast-radius containment
  • weak secret handling or integrity enforcement
  • controls documented by policy but not enforced in code or architecture
  • poor incident investigation support
  • failure paths that disable or bypass protection

Separate confirmed findings from open questions.

6. Deliver a concise report

Default to this structure:

  1. Scope reviewed
  2. Defensive posture summary
  3. Key control points
  4. Coverage gaps and open questions
  5. Recommended defensive improvements

Output Rules

  • Use D3FEND-oriented or defensive-control language when it helps, but stay practical.
  • You do not need official D3FEND artifacts or IDs unless the user explicitly asks for them.
  • Prefer coverage and gap analysis over generic best-practice lists.
  • Be explicit about evidence quality:
    • confirmed when code or architecture directly shows the control
    • inferred when a control likely exists but is not fully proved
    • not verified when runtime, infrastructure, or SOC processes are missing

Search Strategy

Use fast repository discovery first:

  • rg --files for project shape
  • rg for signals such as auth, role, permission, token, secret, key, verify, tls, encrypt, audit, log, alert, trace, policy, allow, deny, retry, limit, validate, sanitize, isolate, internal, admin
  • focus on files that establish controls, trust boundaries, observability, safe defaults, or containment

Example Requests

  • Review this repo using MITRE D3FEND-style defensive coverage analysis.
  • Assess defensive gaps in this architecture.
  • Inspect this codebase and identify where detection, hardening, or containment are weak.