Works with the AI Agents you already use
STRIDE Threat Model and Data Flow Diagram Generator
Evidence-based STRIDE threat model from your repo: Mermaid data flow diagram with trust boundaries, threats traced to file:line, top 5 fixes. LOW RISK, ACTION NEEDED or HIGH RISK.
$7
STRIDE Threat Model and Data Flow Diagram Generator
Example session with this skill installed
Threat model our invoicely repo (Express API, Postgres, Stripe webhooks, an admin panel, Terraform under infra/). It's multi-tenant by org_id. We need something for our SOC 2 auditor and a list of what to fix first.
- Read your context and instructions
- Compiled the stride threat model
Threat model - invoicely @ 4c19a2e
Verdict: HIGH RISK Mode: NEW Method: STRIDE per element on a DFD
Defaults used: no WAF assumed; tenancy detected (org_id on 7 tables)
Data flow diagram (level 0)
flowchart LR
user((Customer browser))
stripe((Stripe))
subgraph TB1["TB1 Internet / edge"]
lb[ALB 443]
end
subgraph TB2["TB2 App"]
api[api - Express]
admin[admin routes]
end
subgraph TB3["TB3 Data"]
db[(postgres)]
end
user -- "F1 HTTPS /api/* [PII, session]" --> lb
lb -- "F2 HTTP :8080 [PII]" --> api
api -- "F3 SQL [PII, payment refs]" --> db
stripe -- "F4 POST /hooks/stripe [payment]" --> api
admin -- "F5 SQL export [PII, all orgs]" --> db
Threats (top 4 of 11)
| ID | Element | STRIDE | Threat | L x I = Sev | Existing mitigation | Gap | Fix |
|---|---|---|---|---|---|---|---|
| T1 | F4 webhook | S, T | Forged event marks invoice paid | 3x3 = CRITICAL | none; searched constructEvent, Stripe-Signature | no signature check, src/routes/hooks.ts:12 | verify signature with the webhook secret before parsing |
| T2 | admin | E | Any logged-in user can call /admin/export | 2x3 = HIGH | requireAuth src/routes/admin.ts:8 | no role check | add role guard, test it |
| T3 | F3 | E, I | Invoice query not scoped by org_id | 2x3 = HIGH | ORM scope on 6 of 7 models | src/models/invoice.ts:41 | add org scope or Postgres RLS |
| T4 | api | D | No rate limit on /api/login | 3x2 = HIGH | none found | brute force, outage | add rate limiter on auth routes |
Assumptions (not verified in code)
- TLS terminates at the ALB (listener 443, infra/lb.tf:22); F2 inside the VPC is plain HTTP
Top 5 fixes
- Verify Stripe webhook signatures - closes T1 - S - src/routes/hooks.ts
- Role guard on admin routes - closes T2 - S - src/routes/admin.ts
- Tenant scope on invoices - closes T3 - M - src/models/invoice.ts
- Rate limit auth routes - closes T4 - S - src/app.ts
- Audit log on export and role changes - closes T7 - M
Design-review aid, not a pen test or certification.
Connects securely to your tools. The creator never sees your data.
About this skill
Threat models are often skipped, or written once from memory and never updated. An AI asked to threat model your app from scratch tends to list generic threats and assume controls you never built. This skill makes the agent read the code and IaC, and require a file and line for every element, every mitigation it credits and every gap it reports.
What it does:
- Finds processes, data stores, external entities and data flows from routes, schemas, API clients, queues and IaC
- Draws a Mermaid data flow diagram with subgraphs as trust boundaries (internet to edge, app to database, service to third-party API, user to admin, tenant to tenant) and numbered, classified flows (PII, credentials, payment, health)
- Applies STRIDE per element, prioritising flows that cross a boundary or carry sensitive data
- Checks for existing mitigations in code (authn and authz middleware, input validation, webhook signatures, TLS, encryption at rest, rate limiting, audit logging, secrets management) and marks each PRESENT, PARTIAL, ABSENT or ASSUMED
- Scores severity as likelihood x impact on a stated 3x3 rule, with the reasoning shown
- Writes docs/security/threat-model.md with DFD, assets, assumptions, threat table and top 5 fixes
- Includes a re-run routine that diffs the model when architecture changes
- Ends with a verdict of LOW RISK, ACTION NEEDED or HIGH RISK
Who it's for: developers, tech leads and security engineers preparing a security design review, scoping a pen test, or collecting threat modeling evidence for SOC 2 or ISO 27001 audits.
Honesty note: this is a design-review aid, not a penetration test, a vulnerability scanner or a certification. It is defensive only and writes no exploit code. It sees what the repository declares at one commit, so controls outside the code (a WAF, network rules set in a console) are listed as assumptions, not counted as mitigations. Mapping the output to specific audit criteria is up to you and your auditor.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.
The model reflects what the code and IaC declare at one commit. Controls outside the repository, runtime configuration and live drift are not visible and are recorded as assumptions. It is not a penetration test, vulnerability scan or certification, and severity scores follow a stated house rule rather than measured likelihood.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 29 days ago
- Software
Git