The OWASP Citizen Development Top 10 is a list of ten security risks that arise when non-specialists build software with low-code, no-code, AI-assisted coding and AI agent tools. It runs from blind trust through logging failures, and it is written for security teams who want to enable this work safely.
What it covers
OWASP describes the project as a guide for organizations that want a culture of security around citizen development. It started with low-code and no-code platforms and now also covers AI-assisted coding and agents. The audience is security professionals first, then the citizen developers themselves, then governance bodies.
A citizen developer here means a business user, such as an analyst or operations lead, who builds an app or automation for their own work without being a professional engineer.
The ten risks in plain words
| ID | Risk | What it means in practice |
|---|---|---|
| CD-SEC-01 | Blind Trust | Accepting platform output, templates or AI code without checking it |
| CD-SEC-02 | Account Impersonation | Apps run on embedded personal or shared identities |
| CD-SEC-03 | Authorization Misuse | Over-broad scopes and long-lived tokens that pile up |
| CD-SEC-04 | Sensitive Data Leakage and Handling Failures | Data flows to places nobody classified or tracked |
| CD-SEC-05 | Authentication and Secure Communication Failures | Plaintext secrets, disabled TLS checks, personal logins |
| CD-SEC-06 | Vulnerable and Untrusted Components | Nested sub-workflows and marketplace parts nobody vetted |
| CD-SEC-07 | Security Misconfiguration | Public endpoints and oversharing from permissive defaults |
| CD-SEC-08 | Injection Handling Failures | Unsanitized input, including platform-specific syntax |
| CD-SEC-09 | Asset Management Failures | Orphaned, unpatched apps outside any inventory |
| CD-SEC-10 | Security Logging and Monitoring Failures | Too little logging, or debug logging that captures secrets |
Where to start
You do not need to tackle all ten at once. Three give the best first return for most engineering organizations.
- Inventory (CD-SEC-09). You cannot judge risk in apps you cannot list.
- Connections (CD-SEC-02, 03 and 05). Most of the damage in OWASP's own scenarios comes from a credential that was too powerful, too shared or too old.
- Data flow (CD-SEC-04). Find which apps move customer data and where it lands.
Blind trust deserves a note. OWASP treats it as foundational, because it amplifies the other nine. The framework it proposes has three actions: enforce secure defaults, extend governance to every citizen-built app, and build security guidance into the tools people already use.
How this differs from the OWASP web Top 10
The web list assumes engineers writing code. This one assumes the builder may not know that a connection stores a token, that a template can hide a script, or that a "public" setting exposes data. The mitigations lean on platform controls, approved component libraries and short training, because you cannot code review a flow nobody told you about.
A one-week plan
- Day one: list the platforms in use and who administers each.
- Day two: export the apps and connections each platform can report.
- Day three: mark apps that touch customer, financial or employee data.
- Day four: flag connections that use a person's identity.
- Day five: pick owners for the top ten risky apps and agree a review date.
Where Safeguard fits
When citizen-built logic lives in repositories, such as custom code blocks or scripts, Safeguard's SBOM and dependency scanning can inventory the packages involved, and the free scan at https://gold.safeguard.sh/scan is a quick way to try it. It does not replace a platform's own admin controls.
Quick check
Is the Citizen Development Top 10 only about low-code platforms?
No. OWASP broadened it to cover AI-assisted coding and AI agents as well as low-code and no-code platforms.
Which risk should we address first?
Start with inventory, because every other control depends on knowing what exists. Then look at connections and data flow.
Is this a compliance standard?
No. It is a risk awareness and guidance project, not a certification or regulation.
Sources
Last reviewed: 19 October 2026.
