Skip to content
Back to skills

Grc Third Party Risk Diagram

ASecurity

Use when creating a draw.io diagram for vendor intake, tiering, questionnaires, security/privacy/legal review, contracting, and ongoing monitoring in a GRC, security, audit, compliance, privacy, cloud, or risk context.

  • 391 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added May 28, 2026
securityrustgobashsecurity

Works with

  • cli

Security analysis

A100/100

Scanned May 28, 2026

npx -y skills add GRCEngClub/claude-grc-engineering --skill grc-third-party-risk-diagram --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Grc Third Party Risk Diagram?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Grc Third Party Risk Diagram
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/grcengclub-grc-third-party-risk-diagram/badge)](https://www.skillsdirectory.com/skills/grcengclub-grc-third-party-risk-diagram)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: grc-third-party-risk-diagram
description: Use when creating a draw.io diagram for vendor intake, tiering, questionnaires, security/privacy/legal review, contracting, and ongoing monitoring in a GRC, security, audit, compliance, privacy, cloud, or risk context.
allowed-tools: Write, Bash, Read, WebFetch
---

# GRC third-party risk diagram

Use this skill to structure the GRC content and visual pattern for vendor intake, tiering, questionnaires, security/privacy/legal review, contracting, and ongoing monitoring. Then use the `drawio` skill to generate the native editable `.drawio` file and optional PNG/SVG/PDF export.

## Common Requests

- new vendor intake
- security questionnaire workflow
- vendor tiering
- DPIA handoff
- continuous vendor monitoring

## Recommended Elements

Include these when relevant:

- business owner
- procurement
- legal
- privacy
- security
- vendor
- approver
- artifacts
- tiering decisions

## Recommended Output Pattern

Produce a Vendor lifecycle, swimlane workflow, or tiering decision tree. Choose a layout that matches the audience:

- Executive: compact lifecycle/capability view with business impact labels.
- Auditor/assessor: explicit evidence, owner, control, cadence, and scope labels.
- Practitioner/engineering: operational systems, data paths, automation, failure/exception paths, and implementation detail.

## draw.io Instructions

1. Load and follow the `drawio` skill.
2. Generate native mxGraphModel XML directly. Do not generate Mermaid as the final artifact.
3. Use descriptive lowercase hyphenated filenames.
4. Include a legend when colors, edge styles, or containers have compliance meaning.
5. Validate XML well-formedness before finalizing.
6. If PNG/SVG/PDF is requested, export with embedded diagram XML when the draw.io CLI is available.

## Visual Conventions

- Blue: systems, platforms, services, and automated collectors.
- Green: implemented controls, approvals, validated evidence, and compliant outcomes.
- Orange/red: risks, findings, exceptions, overdue items, gaps, and failed controls.
- Gray: manual tasks, external parties, optional steps, and out-of-scope areas.
- Dashed containers: audit scope, trust boundaries, authorization boundary, or responsibility boundary.
- Solid edges: primary process or system flow.
- Dashed edges: evidence or attestation flow.
- Dotted edges: optional, manual, exception, or escalation flow.

## Quality Bar

- Make ownership explicit.
- Label regulated data, control IDs, frameworks, and evidence repositories when known.
- Show decision criteria where the process branches.
- Avoid generic boxes like "Compliance" without a role, system, artifact, or action.
- Prefer editable source of truth over screenshots.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.