Add serpdive provider

[edendalexis]

What

Adds serpdive as a provider, mirroring the tavily module. SERPdive returns extracted, answer-ready page content per result instead of snippets, so the snippet field carries real page content. On a public 1,000-question benchmark it won 60.7% of decided quality duels against Tavily's default (basic) search while returning 20.2% fewer tokens (1,001 vs 1,255 per query on average) at the same speed; Tavily's advanced tier was not tested. Free tier, no card.

Full disclosure: I built SERPdive — close freely if you'd rather keep the provider list as is, no hard feelings.

How

• providers/serpdive.ts: bearer auth, POST body { query, max_results } with max_results clamped to the API cap of 10; domain filters fold into the query via appendDomainFilters (the API has no domain parameters); results[].content → snippet
• registered in types.ts, provider-endpoints.ts (https://api.serpdive.com/v1/search), providers.ts, config.ts provider list, package.json keywords, README provider bullet
• tests: baseUrl-override loop entry, buildRequest (bearer + body + domain-filter folding + cap), normalizeResponse

Testing

npm run check (tsgo + biome) clean, npm test 110/110.

---

Summary by cubic

Add serpdive as a web search provider that returns extracted page content in snippet for more answer-ready results. Wired into config, endpoints, providers, tests, and docs (README notes the API key has a free tier).

• New Features
  • Request: POST to https://api.serpdive.com/v1/search with bearer auth; body { query, max_results } (clamped to 10), no model sent; domain filters appended to the query.
  • Normalize: map results[].title/url/content to the common result shape; content → snippet.
  • Wiring & docs: registered in types, provider-endpoints, providers, and config; README updated to note free-tier key; package.json keyword added; tests for request building and response normalization.

^{Written for commit 6e3d869. Summary will update on new commits.}

[edendalexis]

Update while this is open: SERPdive now has a free tier, which changes the calculus for this integration.

model: "krill" is free and unlimited under fair use — no card, no credits to spend, no trial clock. It returns about 700 tokens per search (roughly half what the usual alternatives send), one request at a time, at low priority. The paid models are unchanged.

Why it matters here: the usual objection to adding a search provider is that reviewers and users need a paid key just to try it. That objection is gone — anyone can run this integration end to end for free, including whoever reviews this PR.

Nothing in this PR needs to change for that: the default stays mako, and krill is opt-in via the existing model parameter. Happy to add a line to the docs here if you'd like the free tier called out explicitly.