Fix Autoblog OpenAI fallback by using a valid default text model

[Copilot]

Autoblog generation stopped progressing when Anthropic quota was exhausted because the OpenAI fallback path defaulted to an invalid model name. This change updates the default fallback model so scheduled generation/delivery can continue on OpenAI.

• Backend AI fallback default

  • Updated lib/env.ts to default BACKEND_AI_OPENAI_MODEL to gpt-5-mini (replacing gpt-5.5).

• Operator-facing config parity

  • Updated .env.example to match the runtime default so deployments inherit a valid OpenAI fallback model unless overridden.

• Regression coverage

  • Added tests/contract/env-backend-ai.test.ts to assert the default OpenAI Autoblog model remains gpt-5-mini.

// lib/env.ts
backendAiOpenaiModel:
  process.env.BACKEND_AI_OPENAI_MODEL ?? "gpt-5-mini",

[github-actions]

vu1nz Security Review

0 finding(s) in PR #?

No security issues found.