auth-implementation-patterns — agentic threat model

Not certain from the listing — The underlying foundation model is not specified. The primary threat at this layer is model reprogramming or prompt injection that could force the model to output insecure or backdoored authentication patterns instead of secure-by-default ones.

The agent relies on a static or dynamic knowledge base of auth patterns (JWT, OAuth2, RBAC). The critical threat is knowledge-base poisoning, where an attacker injects flawed or vulnerable code patterns (e.g., weak JWT verification) into the skill's reference repository.

The skill is designed to inject code patterns into an orchestrating agent's implementation context. Framework-level threats include insecure context injection and the risk of the host agent executing unvalidated code snippets generated by this skill.

Not certain from the listing — The hosting environment, sandboxing, and infrastructure are not described. If deployed without isolation, the primary threat is the host agent executing generated code in an unsandboxed environment.

Not certain from the listing — There is no mention of built-in guardrails, logging, or evaluation mechanisms to verify that the injected auth patterns remain secure and free from drift or adversarial manipulation.

The agent directly addresses security and compliance by providing RBAC, SSO, and multi-tenancy patterns. However, there is no evidence of automated compliance auditing (e.g., NIST/ISO alignment) of the generated code itself.

This skill is designed to be consumed by other developer agents. The primary ecosystem threat is cascading vulnerability propagation, where multiple downstream agents consume a poisoned or flawed auth pattern, compromising multiple systems.