https://hackyjs.com/posts/auditing-an-llm-gateway-for-auth-bypass-api-key-exposure-and-tenant-leaks Hacky JS en 2026-10-16T00:00:00.000Z Auditing an LLM Gateway for Auth Bypass, API-Key Exposure, and Tenant Leaks llm-gateway, appsec, auth-bypass, api-security, multi-tenancy https://hackyjs.com/posts/auditing-cloud-hosted-ai-agents-for-runaway-egress-and-quota-abuse Hacky JS en 2026-10-16T00:00:00.000Z Auditing Cloud-Hosted AI Agents for Runaway Egress and Quota Abuse ai-agents, cloud-security, auditing, egress-controls, quota-enforcement https://hackyjs.com/posts/using-eval-gates-to-decide-when-a-cheap-open-weight-model-is-good-enough-for-an-agent-task Hacky JS en 2026-10-16T00:00:00.000Z Using Eval Gates to Decide When a Cheap Open-Weight Model Is Good Enough for an Agent Task ai-agents, llm-evaluation, inference-costs, open-weight-models, evals https://hackyjs.com/posts/why-single-provider-ai-clients-break-under-load-and-how-to-route-around-them Hacky JS en 2026-10-16T00:00:00.000Z Why Single-Provider AI Clients Break Under Load and How to Route Around Them ai-apis, reliability, multi-provider-routing, graceful-degradation, observability