Add a new LLM provider adapter to Primer (e.g. Gemini, Mistral, Cohere). Plugs into the registry / dispatcher pattern so the pipeline, chat, and analytics surface the new provider without touching the calling code. Use when adding a provider whose models should appear in the per-use-case AI model picker.
Add a new step to the briefing-generation pipeline (the 9-step flow in services/briefing-generator.ts). Covers checkpointing, cancellation, error isolation, retry, and timing recording. Use when inserting a new transformation between fetch / extract / generate stages.
Add a new Hono API route to the Primer worker. Covers request / response shape, admin gating, error handling, and the frontend apiGet / apiPost / apiPatch / apiDelete pairing. Use when wiring a new endpoint or moving an existing one.
Add a new TTS (text-to-speech) provider adapter (e.g. Azure, Polly, PlayHT). Plugs into the registry / dispatcher pattern so the per-operation voice picker surfaces the new provider's voices. Use when adding a voice provider whose voices should appear in the Voice settings panel.
Add or modify auth providers in Primer's identity-resolution layer. Covers the AuthProvider interface, the email allowlist contract, the fail-closed factory, and the source-text contract tests every provider must add. Use when adding support for a new auth proxy (oauth2-proxy, Pomerium, Tailscale Serve, nginx + OIDC), modifying an existing provider, or working in src/worker/middleware/auth/.
Create or modify source providers that feed data into Primer briefings. Covers the full stack: integration client, source provider, registry, settings manifest, and optional custom frontend panel. Use when adding a new data source, modifying an existing source's filters or behavior, or working in src/worker/sources/.