security

Overview of security skill

What the security skill does

The security skill creates security architecture diagrams in PlantUML with AWS security stencils for identity, encryption, firewalling, compliance, and threat detection. It is built for readers who need a fast way to turn a security design into a diagram that is clear enough for review, audit, or documentation.

Best-fit use cases

Use this security skill when your real job is to explain IAM flows, zero-trust boundaries, encryption pipelines, compliance monitoring, or detection-to-response paths. It is especially useful for Security Audit, architecture reviews, and docs that need a precise security model instead of a generic cloud diagram.

Why it is worth installing

The main value is the stencil-driven workflow: the skill pushes you toward security-specific icons, trust boundaries, and directional flows that make the diagram easier to read and harder to mislabel. It is a better fit than a plain prompt when you need consistency across repeated security diagrams.

How to Use security skill

Install and open the right files

Install the security skill with npx skills add markdown-viewer/skills --skill security. Then read SKILL.md first, followed by the most relevant examples such as examples/iam-authn.md, examples/zero-trust.md, examples/compliance-audit.md, and examples/threat-detection.md. Those files show the intended pattern faster than browsing the whole repo.

Give the skill a complete security goal

Strong inputs describe the security job, not just the diagram title. For example, ask for “a zero-trust access flow for employees and contractors with SSO, STS role assumption, network boundary checks, and audit logging” instead of “draw security architecture.” The skill works best when you provide actors, trust boundaries, security controls, and the outcome you want the viewer to understand.

Use a workflow the skill can follow

A good security usage flow is: define the security objective, list the trust zones, choose the identity and protection components, then connect them with the correct flow direction. If you are mapping Security Audit, specify the control sources, evidence store, remediation path, and reporting target so the diagram shows both control and proof.

Read the syntax rules before generating

The repository is strict about PlantUML formatting: use @startuml / @enduml, keep access flows left to right, and use ```plantuml or ```puml fences only. It also expects rectangle trust boundaries and mxgraph.aws4.* stencil syntax, so avoid inventing custom shapes unless the diagram needs them.

security skill FAQ

Is this security skill for general cloud diagrams?

No. The repo explicitly says not to use it for general cloud infrastructure; use a cloud-focused skill instead. This security skill is for diagrams where the main story is identity, encryption, perimeter defense, compliance, or detection.

Is it useful for Security Audit work?

Yes, especially when you need to show evidence flow, policy checks, centralized logs, and remediation loops. For Security Audit, the strongest diagrams usually combine Config, Audit Manager, CloudTrail, Security Hub, and storage or reporting outputs so the audit path is visible end to end.

Can a beginner use it without knowing PlantUML well?

Yes, but only if the request includes a clear architecture narrative. Beginners get better results when they name the actors, trust zones, and controls they already know, then let the skill translate that into a security diagram. A vague prompt usually produces a vague diagram.

When should I not use it?

Do not use it if you only need a high-level system sketch, product overview, or non-security software model. If the main decision is compute topology, service dependencies, or generic UML structure, another skill will be a better fit.

How to Improve security skill

Start with the control objective

The best security outputs come from a single clear objective: “prevent lateral movement,” “enforce least privilege,” “prove compliance evidence,” or “detect and respond to threats.” That objective helps the skill choose the right icons and the right level of detail for Security Audit or other review contexts.

Provide concrete entities and boundaries

Give explicit inputs such as user type, identity provider, protected resource, audit sink, and enforcement points. For example, “employees via AD, customers via Cognito, STS for temporary credentials, API Gateway, and CloudTrail into centralized log archive” is much stronger than “auth flow diagram.”

Avoid mixing too many security stories

A common failure mode is combining IAM, encryption, network security, compliance, and incident response into one overloaded diagram. If the first output feels crowded, split the request into separate diagrams and keep one security storyline per diagram so the flow remains readable.

Iterate with missing pieces, not style-only tweaks

If the result is structurally right but incomplete, ask for the missing control or evidence path rather than just “make it better.” Useful follow-ups are: add trust boundaries, show async audit flows, insert encryption at rest, or replace a generic firewall with the correct AWS security stencil for the use case.