@koda-sl/baker-bridge 0.172.0 → 0.173.0

package/README.md

@@ -45,13 +45,13 @@ await server.start();
 | `ANTHROPIC_API_KEY` | Anthropic API key (required by the Agent SDK) |
 | `ANTHROPIC_BASE_URL` / `ANTHROPIC_AUTH_TOKEN` | OpenRouter routing (base URL + per-org key) injected into the sandbox. Forwarded to the Agent SDK subprocess so the agent can reach its model |
 | `OPENROUTER_API_KEY` | The same per-org key under its own name. The bridge reads the user's and the agent's words with it before letting two kinds of call through: a question asking the user to pick an avatar's face (the casting picker shows real portraits), and an avatar accent in a language nobody asked for. Unset or unreachable, both calls go through unchecked |
-| `CLAUDE_CODE_EXTRA_BODY` | Optional JSON merged by the CLI into every model request body. The bridge sets/extends it per turn with a per-thread `session_id` so OpenRouter sticky-routes the whole conversation to one provider instance and its instance-scoped prompt cache is re-read across the turn's many tool-call requests (see [Prompt-cache pinning](#prompt-cache-pinning)). It also defaults `provider` to Bedrock first (see [Bedrock is the preferred provider](#bedrock-is-the-preferred-provider)). Any inherited value is merged; our `session_id` and the `web`-plugin opt-out win (see [OpenRouter web plugin is forced off](#openrouter-web-plugin-is-forced-off)). An inherited `provider` replaces the Bedrock default |
+| `CLAUDE_CODE_EXTRA_BODY` | Optional JSON merged by the CLI into every model request body. The bridge sets/extends it per turn with a per-thread `session_id` so OpenRouter sticky-routes the whole conversation to one provider instance and its instance-scoped prompt cache is re-read across the turn's many tool-call requests (see [Prompt-cache pinning](#prompt-cache-pinning)). It also pins `provider` to our own Bedrock, with no fallback (see [Claude runs on our own Bedrock only](#claude-runs-on-our-own-bedrock-only)). Any inherited value is merged; our `provider`, `session_id` and the `web`-plugin opt-out win (see [OpenRouter web plugin is forced off](#openrouter-web-plugin-is-forced-off)) |
 | `BAKER_ENVIRONMENT`, `BAKER_DEPLOYMENT`, `BAKER_ORGANIZATION_ID` | Optional. Where the Runtime runs (`production`, `preview`, `development`), the backend deployment's name and the organization. With `BAKER_RUNTIME_COMPANY_ID`, `BAKER_CHAT_ID` and the turn's sender they tag every OpenRouter request the agent and the bridge make — `HTTP-Referer` / `X-OpenRouter-Title` naming the app (`https://agent.<environment>.withbaker.com`), `x-session-id`, and in the body `user` (`<feature> | company:<id> | chat:<id> | user:<id>`, what OpenRouter shows per request) plus `trace` — so a request without the tag was not sent by Baker. The agent gets the headers through `ANTHROPIC_CUSTOM_HEADERS` (merged after any inherited value) and the body fields through `CLAUDE_CODE_EXTRA_BODY`; a Session review is tagged `chatReview.session`, a task `builder.task`, a plan `agent.plan`, a chat turn `agent.turn` |
 | `BAKER_CHAT_ID` | Present only when running inside a chat sandbox. When set, the bridge relocates Claude Code auto-memory into the cloned repo (`.claude/memory`) — see [Session journal](#session-journal-agent-memory) |
 | `BAKER_ACTING_USER_ID` | Injected into the Agent SDK subprocess env per turn (when the turn's sender is known) so in-sandbox `baker` CLI calls can scope work to the acting user — e.g. `baker mcp add --scope user` |
 | `CLAUDE_CODE_ENABLE_TASKS` | Set to `0` on every Agent SDK subprocess. Turns off the CLI's built-in task backlog and its `task_reminder` — see [Session surface](#session-surface) |
-| `CLAUDE_CODE_DISABLE_TERMINAL_TITLE` | Set to `1` on every Agent SDK subprocess. The CLI's session-title request carries a structured output OpenRouter cannot route to Bedrock — see [Bedrock is the preferred provider](#bedrock-is-the-preferred-provider) |
-| `CLAUDE_CODE_MODEL_CAPABILITIES` | Set to `-per_turn_effort` on every Agent SDK subprocess so no request carries a per-message effort, which OpenRouter cannot route to Bedrock — see [Bedrock is the preferred provider](#bedrock-is-the-preferred-provider) |
+| `CLAUDE_CODE_DISABLE_TERMINAL_TITLE` | Set to `1` on every Agent SDK subprocess. The CLI's session-title request carries a structured output OpenRouter cannot route to Bedrock — see [Claude runs on our own Bedrock only](#claude-runs-on-our-own-bedrock-only) |
+| `CLAUDE_CODE_MODEL_CAPABILITIES` | Set to `-per_turn_effort` on every Agent SDK subprocess so no request carries a per-message effort, which OpenRouter cannot route to Bedrock — see [Claude runs on our own Bedrock only](#claude-runs-on-our-own-bedrock-only) |
 | `NO_COLOR` | Set to `1` on every Agent SDK subprocess so `baker --help` output carries no ANSI escapes — see [Session surface](#session-surface) |
 | `BAKER_AGENT_MODEL` | Optional. Runs every turn in this runtime on the named model instead of the default, and is forwarded to subagents as `CLAUDE_CODE_SUBAGENT_MODEL`. Ignored unless it names a model with a known context window — see [Model selection](#model-selection) |
 | `CLAUDE_CODE_SUBAGENT_MODEL`, `CLAUDE_CODE_SUBAGENT_MODEL_FORCE` | Set on the Agent SDK subprocess only when `BAKER_AGENT_MODEL` is set. Together they override the `model:` each subagent pins in its own frontmatter |
@@ -200,31 +200,28 @@ re-paid full input price on the same ~30k-token prefix.
 The fix (`openrouter-session.ts`): the bridge injects a stable per-thread
 `session_id` via `CLAUDE_CODE_EXTRA_BODY`, which the CLI merges into every model
 request body. That pins the whole conversation to its warm provider instance "even
-if the opening messages change". `session_id` alone does not disable fallbacks, so
-provider resilience is preserved. See OpenRouter → Prompt Caching (best practices). A manual
+if the opening messages change". See OpenRouter → Prompt Caching (best practices). A manual
 `provider.order` bypasses this sticky routing — see
-[Bedrock is the preferred provider](#bedrock-is-the-preferred-provider) for how the
-cache stays warm while the Bedrock default is set.
+[Claude runs on our own Bedrock only](#claude-runs-on-our-own-bedrock-only) for how the
+cache stays warm while the Bedrock pin is set.
 
-## Bedrock is the preferred provider
+## Claude runs on our own Bedrock only
 
 The same `CLAUDE_CODE_EXTRA_BODY` value carries a `provider` with
-`allow_fallbacks: true` and this `order`:
+`allow_fallbacks: false` and these endpoints as both `only` and `order`:
 
 ```json
 ["amazon-bedrock", "amazon-bedrock/eu-west-1", "amazon-bedrock/us-east-1"]
 ```
 
-Without it, OpenRouter load-balances Claude across Anthropic, Vertex, Azure and
-Bedrock by price and uptime, and the `session_id` pin then holds the conversation
-on whichever it picked first. With it, every request goes to Bedrock first, then
-to the two regional Bedrock endpoints, and only moves to another provider when all
-three fail. The regions are there because Bedrock's failure mode is AWS capacity,
-not our quota: on weekday mornings the default endpoint answers `503` to most Opus
-5.5 requests for hours, and AWS's advice for a `503` is another region. All three
-run on our own Bedrock key; the providers after them bill OpenRouter credits. An
-inherited `provider` (set in the sandbox's `CLAUDE_CODE_EXTRA_BODY`) replaces this
-default as an operator override.
+Every Claude request is served by Amazon Bedrock on Baker's own key (OpenRouter
+BYOK), or the turn fails. Anthropic, Vertex, Azure and OpenRouter's own capacity
+are paid from OpenRouter credits, and an error is preferred to that charge. The
+regions are there because Bedrock's failure mode is AWS capacity, not our quota:
+on weekday mornings the default endpoint answers `503` to most Opus 5.5 requests
+for hours, and AWS's advice for a `503` is another region. All three run on our
+own key. An inherited `provider` (set in the sandbox's `CLAUDE_CODE_EXTRA_BODY`)
+is replaced: the pin is not an operator override.
 
 A manual `order` bypasses OpenRouter's sticky routing, so while it is set the
 `session_id` pin above is inert and the `order` is what keeps the cache warm. The
@@ -237,19 +234,15 @@ and each request retries the default endpoint first, so a cache warmed further
 down the `order` is not reused while Bedrock is flapping. During a capacity
 window every request also waits for the default endpoint's `503` (about 7 s)
 before moving on.
-The `session_id` pin still applies whenever an inherited `provider` drops the
-`order`.
-
-The preference only holds for a request Bedrock can serve, and OpenRouter drops
-Bedrock from the route without an error when it can't. From Opus 5.5 the CLI
+The pin only works for a request Bedrock can serve; one it can't now fails
+instead of moving to another provider. From Opus 5.5 the CLI
 restates the effort on its mid-conversation `role: "system"` message
 (`output_config.effort`, beta `per-turn-control`), which no Bedrock endpoint
-supports — so the day Opus 5.5 shipped, 52 of ~4,600 of its requests reached
-Bedrock and the rest were billed to OpenRouter credits. The bridge sets
+supports — so the day Opus 5.5 shipped, before the pin, 52 of ~4,600 of its
+requests reached Bedrock and the rest were billed to OpenRouter credits. The bridge sets
 `CLAUDE_CODE_MODEL_CAPABILITIES=-per_turn_effort` to stop that; the request-level
-`effort` still applies. A request with a structured-output schema is served
-elsewhere, because Bedrock lacks structured outputs on OpenRouter; the CLI's one
-such request, naming the session, is never sent
+`effort` still applies. The CLI's request naming the session is never
+sent
 (`CLAUDE_CODE_DISABLE_TERMINAL_TITLE=1`), since nothing in Baker reads the name.
 
 ## OpenRouter web plugin is forced off

package/dist/hono/bridge-job.test.js

@@ -1279,13 +1279,14 @@ describe("BridgeJobExecutor", () => {
             BAKER_THREAD_ID: "thread_env",
             BAKER_TURN_ID: "turn_env",
         });
-        // Guards the Bedrock preference, the prompt-cache pin and the OpenRouter
+        // Guards the Bedrock-only pin, the prompt-cache pin and the OpenRouter
         // web-plugin opt-out at their integration point: all must actually reach the
         // subprocess body via CLAUDE_CODE_EXTRA_BODY.
         expect(JSON.parse(env?.CLAUDE_CODE_EXTRA_BODY ?? "{}")).toEqual({
             provider: {
+                only: ["amazon-bedrock", "amazon-bedrock/eu-west-1", "amazon-bedrock/us-east-1"],
                 order: ["amazon-bedrock", "amazon-bedrock/eu-west-1", "amazon-bedrock/us-east-1"],
-                allow_fallbacks: true,
+                allow_fallbacks: false,
             },
             session_id: "thread_env",
             user: "agent.turn | company:company_env | chat:chat_env",