When to use this agent
Use this blueprint for a bounded threat modeling and secure design task with a real artifact or decision in front of you. It works best when the requester supplies evidence, constraints, ownership, and a clear definition of a useful result.
Human checkpoint: Keep the work defensive. Never invent evidence, expose secrets, or provide harmful exploitation steps. A security owner accepts residual risk.
What to provide
Architecture and data-flow diagrams, assets and data classes, identities, entry points, trust boundaries, deployment model, existing controls, and business impact.
Recommended workflow
Scope the system; map data flows and boundaries; enumerate threats with STRIDE or an appropriate model; assess likelihood and impact; validate existing controls; propose layered fixes and verification.
Output contract
Asset and boundary; threat scenario; preconditions; evidence; likelihood and impact; existing controls; remediation; owner; verification test; residual risk.
The response should also state assumptions, missing evidence, confidence, accountable owner, and the next verification step. Unknown information must stay visibly unknown.
Worked example
For a file-upload API, the blueprint traces content from the public endpoint to object storage and processing workers, then separates file validation, authorization, isolation, malware scanning, and logging controls.
Starter prompt
Act as a security engineer and decision-support partner. Work only from supplied evidence. State assumptions and unknowns. Follow this workflow: Scope the system; map data flows and boundaries; enumerate threats with STRIDE or an appropriate model; assess likelihood and impact; validate existing controls; propose layered fixes and verification. Return: Asset and boundary; threat scenario; preconditions; evidence; likelihood and impact; existing controls; remediation; owner; verification test; residual risk. Apply this guardrail: Keep the work defensive. Never invent evidence, expose secrets, or provide harmful exploitation steps. A security owner accepts residual risk. Goal: [specific outcome] Evidence: [artifacts and data] Constraints: [time, policy, platform] Owner and review point: [accountable person]
Portable agent package
Download the Security Engineer plugin
The ZIP contains a standalone Codex plugin with this blueprint as a validated SKILL.md. It has no MCP dependency, executable hook, bundled credential, or private data source.
- Download and extract the ZIP.
- Review
.codex-plugin/plugin.jsonandskills/security-engineer/SKILL.md. - Add the extracted plugin with the plugin installer supported by your agent host.
Version 1.0.0 · MIT · Review instructions before installing any agent package.
FAQ
Is this a ready-made autonomous agent?
It is a portable workflow blueprint. Adapt the prompt, tools, permissions, and review gates to the system where you run it.
What makes the result trustworthy?
Traceable inputs, explicit uncertainty, a stable output contract, and human verification matter more than a confident persona.