@arnilo/prism-provider-opencode-go

OpenCode Go provider package for Prism (official docs).

import {
  createOpenCodeGoProviderPackage,
  listOpenCodeGoModels,
} from "@arnilo/prism-provider-opencode-go";

api.registerProviderPackage(createOpenCodeGoProviderPackage({ apiKey: "fake-opencode-key" }));

const live = await listOpenCodeGoModels({ apiKey: "fake-opencode-key" });
api.registerProviderPackage(createOpenCodeGoProviderPackage({ apiKey: "fake-opencode-key", models: live }));

Exports:

• createOpenCodeGoProviderPackage() / createOpenCodeGoProvider()
• openCodeGoModels (docs-verified featured aliases)
• listOpenCodeGoModels() / mapOpenCodeGoModel() / defineOpenCodeGoModel() / routeForOpenCodeGoModel()
• OPENCODE_GO_DEFAULT_BASE_URL (https://opencode.ai/zen/go/v1)
• Thinking helpers: openCodeGoThinking, openCodeGoReasoningEffort, openCodeGoReasoning, openCodeGoPreserveThinking, stripOpenCodeGoOwnedCompat
• Cache helpers: opencodeSessionId, opencodeOwnedHeaders, applyOpencodeAnthropicCacheControl

Security defaults:

• No network calls during import, setup, build, or default tests.
• No automatic environment, file, keychain, or shell credential lookup.
• API keys are resolved per request from caller-supplied values or resolvers and redacted from errors (including discovery).
• Live tests must stay opt-in behind PRISM_LIVE_PROVIDER_TESTS=1 plus OPENCODE_API_KEY.

Cache, session, thinking:

• Default base URL: https://opencode.ai/zen/go/v1 (Chat Completions + Anthropic Messages + /models).
• x-opencode-session from cacheKey ?? sessionId, sanitized + clamped to 128 chars.
• Anthropic route (minimax-*, qwen*) applies cache_control only to selected breakpoints ("long" → ttl: "1h").
• OpenAI route (Grok/GLM/Kimi/MiMo/DeepSeek) sends no Anthropic cache_control; implicit cache + reasoning_content preserve.
• Per-route usage mapping unchanged.
• Provider-owned headers (content-type, x-opencode-session, authorization) win over caller headers.