Build security into software decisions before they become expensive defects.
\nThis skill helps an agent turn \"make it secure\" into concrete design choices: what must be protected, who may do what, where trust changes, and how a team can verify the result. It covers everyday engineering work such as APIs, credentials, dependencies, multi-tenant services, release artifacts, and AI features without assuming a cloud provider or framework.
\nInstead of treating a checklist or scanner as a security guarantee, the workflow asks for evidence, assumptions, misuse cases, and accountable residual-risk decisions. It complements an assessment skill: use this one while building and changing software, then assess or scan separately when that is the actual task.
\n| Path | \nWhat it provides | \n
|---|---|
SKILL.md | \nA five-phase, prevention-oriented workflow and routing guide. | \n
references/source-index.md | \nVersion-pinned primary sources and the decisions they inform. | \n
references/ | \nFocused guidance for threat modeling, controls, review, release, incident learning, AI systems, and multi-tenant boundaries. | \n
templates/ | \nAdaptable threat-model, acceptance-criteria, and review-checklist starting points. | \n
Install or expose this directory through your Agent Skills-compatible client. Then ask, for example:
\nThreat-model this multi-tenant document API before implementation.\nThe resulting model should name assets, boundaries, assumptions, abuse cases, mitigations, evidence, and residual risk rather than simply declare the API secure.
\nNo runtime dependency, API key, or platform-specific tool is required. The included sources are decision aids; adopt the controls that apply to your organization, contract, regulator, and threat model.
\nFor end-to-end SaaS tenant semantics and placement decisions, use\nmulti-tenant-saas-architecture. For quantitative tenant capacity and cost\nevidence, use capacity-and-cost-engineering; for privacy lifecycle artifacts,\nuse privacy-engineering.