@theokit/sdk 5.10.0 → 5.11.1

package/CHANGELOG.md

@@ -1,5 +1,87 @@
 # Changelog
 
+## 5.11.1
+
+### Patch Changes
+
+- [#740](https://github.com/usetheokit/theokit-sdk/pull/740) [`f5ee221`](https://github.com/usetheokit/theokit-sdk/commit/f5ee2218c3fc7dfba0b87de64c7472b8fd7184e8) Thanks [@usetheodev](https://github.com/usetheodev)! - `run.cancel()` on a local run now aborts the provider request. Until now it only set the status to `cancelled`: the LLM request kept streaming, `onDelta` kept receiving updates, and the provider billed the whole generation. The run now aborts the request, stops waiting for in-flight tool calls and passes the abort to custom tool handlers through `ctx.signal`, and delivers no update to `onDelta` once it is terminal. Shell commands and MCP tool calls receive no abort and keep running until they finish. A run cancelled while its MCP servers start up runs no turn, and still closes the MCP clients it owns. A cancelled turn reports no `usage` or `cost`: its result resolves when you cancel, before the turn's usage is collected. A `BudgetTracker` receives no tokens for a cancelled turn either, although the provider bills what it generated before the abort. A failed `close()` of a run-owned MCP client is now reported as a diagnostic naming the server, where it used to be dropped. Once a run is cancelled, `stream()`, `events()` and `conversation()` stop growing: the `[aborted]` message and the conversation the turn returns while it winds down are no longer added after `wait()` has resolved.
+
+- [#744](https://github.com/usetheokit/theokit-sdk/pull/744) [`afb44b1`](https://github.com/usetheokit/theokit-sdk/commit/afb44b14153b15ddf482e78b89b4541f7c99836f) Thanks [@usetheodev](https://github.com/usetheodev)! - A provider that answers HTTP 404 because it does not serve the requested model now fails the run with a `ConfigurationError` whose message names the model. Until now OpenRouter's 404 for a retired model id (`No endpoints found for deepseek/deepseek-r1.`) became an `UnknownAgentError` with code `openrouter_unknown`, and neither the error nor `result.error.message` said which model was refused. The run now reports `result.error.code` `model_unavailable`, the error's `code` is `<provider>_model_unavailable`, and the message ends with the provider's own text when the provider sends one. OpenAI's 404 with code `model_not_found` gets the same class, at every OpenAI-compatible call site. OpenRouter's other "No endpoints found" 404s, for a request your data policy or the model's tool support rules out, report the same `model_unavailable` code; the provider's text in the message tells them apart. Any other 404, such as a wrong path, is still an `UnknownAgentError` with its old message.
+
+- [#743](https://github.com/usetheokit/theokit-sdk/pull/743) [`a87a239`](https://github.com/usetheokit/theokit-sdk/commit/a87a23961f2804e37575b12a8f7a5bb59f0b8bd7) Thanks [@usetheodev](https://github.com/usetheodev)! - The `thinking-completed` update now arrives when the reasoning ends: before the first `text-delta`
+  of a turn that reasons and then answers. It used to arrive after the whole turn's stream, so a UI
+  that closed its reasoning block on it kept the block open until the answer had finished.
+  
+  `thinkingDurationMs` on that update, and `thinking_duration_ms` on the replayed `thinking` message,
+  are now the reasoning span: the time of the last reasoning delta minus the time of the first. They
+  used to run until the end of the stream and so included the answer. In a live OpenRouter recording
+  the reported value was 19076 ms for 8282 ms of reasoning. Reasoning that arrives in a single delta,
+  with no delta after it, reports 0.
+  
+  A turn that reasons and then calls a tool, or that ends without text, still sends the update at the
+  end of the stream, with the same duration rule. Each turn with reasoning still sends exactly one
+  `thinking-completed`; reasoning that arrives after the answer started is shown and replayed but
+  sends no second one.
+
+## 5.11.0
+
+### Minor Changes
+
+- [#734](https://github.com/usetheokit/theokit-sdk/pull/734) [`1f04344`](https://github.com/usetheokit/theokit-sdk/commit/1f0434491b706f7ea97f53d998fc33b300af1c33) Thanks [@usetheodev](https://github.com/usetheodev)! - Prompt caching now turns on for Anthropic models, and the cache counters they report reach the
+  turn's usage.
+  
+  For a model id that starts with `anthropic/` or `claude` (after an optional `openrouter/`,
+  `vertex/` or `bedrock/` prefix),
+  the agent loop sends the system prompt as text blocks: the configured prompt first, marked with
+  `cache_control: { type: "ephemeral" }`, then the `<active-memory>` block when memory adds one. The
+  Anthropic client sends the blocks as they are, and the OpenRouter client sends them as `content`
+  parts of the `system` message, with `cache_control` on the cacheable part. Every other provider
+  still receives one string. Bedrock ids (`anthropic.claude-…`) are not matched,
+  and their bodies do not change. Vertex Anthropic ids (`vertex/anthropic/claude-…`) are matched
+  and get the same blocks; no live Vertex response was measured for this change.
+  
+  Things a caller can observe:
+  
+  - **Billing.** Anthropic charges cache writes at 1.25 times the input price and cache reads at 0.1
+    times. A run with a single turn whose system prompt is above the model's caching minimum pays
+    the write premium and never reads the prefix back.
+  - **Order.** For these models the memory block now comes after the configured prompt instead of
+    before it. On a client that joins the blocks into one string, the text is the same and the order
+    is reversed.
+  - **What still misses the cache.** The cached prefix is the whole assembled prompt, so `<memory>`
+    facts, resolved skills or context that change between turns change block 1 and miss.
+  - **Counters.** The Anthropic streaming client reads `cache_read_input_tokens` and
+    `cache_creation_input_tokens` from `message_start`, where Anthropic reports them; a value in
+    `message_delta` still wins when it is above zero, and a zero there leaves the `message_start`
+    value in place. The non-streaming parser shared by the Bedrock and Vertex Anthropic clients now
+    returns both counters too.
+  - **Cost on the chat-completions client.** `prompt_tokens` there already includes the cached
+    slice, and the client used to report it whole as `inputTokens` beside `cacheReadTokens`, so the
+    estimated cost charged every cached token twice (about 8 times too high on a cached
+    OpenRouter Anthropic turn). `inputTokens` is now the uncached portion, as on the Anthropic
+    client, and `prompt_tokens_details.cache_write_tokens` is read into `cacheWriteTokens`. A
+    native OpenAI chat-completions run with a cache hit reports a lower `inputTokens` and a lower
+    cost than before. When a usage object carries both the nested `prompt_tokens_details` counts and the
+    Anthropic-style top-level ones, the nested count wins and a top-level count fills a nested
+    zero, the rule `normalizeUsage` already applied; the stream client used to let the top-level
+    field win.
+  - **Budget trackers.** A `BudgetTracker`'s `input` event now carries the whole prompt:
+    `inputTokens` plus `cacheReadTokens` and `cacheWriteTokens`. On the chat-completions client this
+    keeps the count a tracker saw before `inputTokens` became the uncached portion. On the Anthropic
+    clients, which always reported the uncached portion, a cached turn now counts its cached tokens
+    against `maxTokens`, and `createUsdBudgetTracker` prices them at the input rate, so a cost cap
+    trips earlier than the provider's bill would.
+  - **Input tokens on the Anthropic stream.** `input_tokens` is read from `message_start` too, so a
+    stream whose `message_delta` carries only `output_tokens` reports its uncached input instead of
+    none. A value above zero in `message_delta` still replaces it; a `null` or a zero there leaves
+    the `message_start` count.
+  - **Malformed Anthropic frames.** A `content_block_start`, `content_block_delta` or
+    `message_delta` frame that is not valid JSON now fails the turn with a retryable `NetworkError`
+    (`code: "stream_malformed_frame"`) instead of being skipped, which finished the turn with text
+    or tool input missing. Any other malformed frame is still skipped and reported on the
+    diagnostics channel. The error message and the diagnostic line name the event and the frame's
+    length, never its body, which can hold model output or tool input.
+
 ## 5.10.0
 
 ### Minor Changes

package/dist/{agent-XTLAVVYC.js → agent-DFV52V6E.js}

@@ -1,19 +1,19 @@
-export { Agent } from './chunk-RTE2DD4S.js';
+export { Agent } from './chunk-MHIBYKWA.js';
 import './chunk-XKMVJ443.js';
-import './chunk-EXI7EHDN.js';
+import './chunk-Q34EBNYG.js';
 import './chunk-K2BQQ445.js';
 import './chunk-DPY2M74T.js';
 import './chunk-ZHWKT2NI.js';
 import './chunk-2SFBB54R.js';
-import './chunk-HI3W37FY.js';
+import './chunk-U27XHOEF.js';
 import './chunk-H73MEMQB.js';
 import './chunk-OR6XIWLB.js';
 import './chunk-X3LOJJMO.js';
 import './chunk-IZDDSLL2.js';
-import './chunk-EVH3OUMY.js';
+import './chunk-BBJ77S62.js';
 import './chunk-SUVCHX4L.js';
 import './chunk-2XC6KDJM.js';
-import './chunk-TE6XVTOA.js';
+import './chunk-JLIRFWUG.js';
 import './chunk-24MGQDK3.js';
 import './chunk-XERM7PDQ.js';
 import './chunk-HKCKOAIO.js';
@@ -47,6 +47,7 @@ import './chunk-ZDYDBXSD.js';
 import './chunk-X7EUUHXU.js';
 import './chunk-MQGSOKR3.js';
 import './chunk-GMRFAX4I.js';
+import './chunk-LZJXEPEQ.js';
 import './chunk-R7WIIPUR.js';
 import './chunk-I2ORT6QD.js';
 import './chunk-SVCPJRUB.js';
@@ -57,5 +58,5 @@ import './chunk-H3CVJSFC.js';
 import './chunk-Q7GYMXVC.js';
 import './chunk-B5MHRQJZ.js';
 import './chunk-CZJ6Q7CW.js';
-//# sourceMappingURL=agent-XTLAVVYC.js.map
-//# sourceMappingURL=agent-XTLAVVYC.js.map
+//# sourceMappingURL=agent-DFV52V6E.js.map
+//# sourceMappingURL=agent-DFV52V6E.js.map

package/dist/{agent-XTLAVVYC.js.map → agent-DFV52V6E.js.map}

@@ -1 +1 @@
-{"version":3,"sources":[],"names":[],"mappings":"","file":"agent-XTLAVVYC.js"}
+{"version":3,"sources":[],"names":[],"mappings":"","file":"agent-DFV52V6E.js"}

package/dist/{agent-TLRURB7W.cjs → agent-OKP4LHG2.cjs}

@@ -1,21 +1,21 @@
 'use strict';
 
-var chunkCVEQ5366_cjs = require('./chunk-CVEQ5366.cjs');
+var chunk35KBC4BX_cjs = require('./chunk-35KBC4BX.cjs');
 require('./chunk-J6GXS5A2.cjs');
-require('./chunk-FR6COQXZ.cjs');
+require('./chunk-3JANYJYS.cjs');
 require('./chunk-BUUUWQMB.cjs');
 require('./chunk-L3C6CCBG.cjs');
 require('./chunk-KIKKTW5O.cjs');
 require('./chunk-BV2MWEMV.cjs');
-require('./chunk-YEHQ3SIC.cjs');
+require('./chunk-JV7JA7BM.cjs');
 require('./chunk-GHX4P3V2.cjs');
 require('./chunk-TMKTSYDS.cjs');
 require('./chunk-ZD66GFH5.cjs');
 require('./chunk-NQVWSSX2.cjs');
-require('./chunk-QY3CIMUV.cjs');
+require('./chunk-P5N3Q2NH.cjs');
 require('./chunk-CQYBJOB2.cjs');
 require('./chunk-YJNYDJ2T.cjs');
-require('./chunk-WPO3VCY4.cjs');
+require('./chunk-N7KS3TIR.cjs');
 require('./chunk-QOGNZMHD.cjs');
 require('./chunk-ZAM7BJXJ.cjs');
 require('./chunk-2ZPEDVLM.cjs');
@@ -49,6 +49,7 @@ require('./chunk-C7GKRGKN.cjs');
 require('./chunk-TY56BKSK.cjs');
 require('./chunk-MPR2ISL2.cjs');
 require('./chunk-J5623RIV.cjs');
+require('./chunk-NWS2CS3Y.cjs');
 require('./chunk-3EE6LVWT.cjs');
 require('./chunk-SGC7LSHF.cjs');
 require('./chunk-HU666T3Q.cjs');
@@ -64,7 +65,7 @@ require('./chunk-6LHQPOMI.cjs');
 
 Object.defineProperty(exports, "Agent", {
   enumerable: true,
-  get: function () { return chunkCVEQ5366_cjs.Agent; }
+  get: function () { return chunk35KBC4BX_cjs.Agent; }
 });
-//# sourceMappingURL=agent-TLRURB7W.cjs.map
-//# sourceMappingURL=agent-TLRURB7W.cjs.map
+//# sourceMappingURL=agent-OKP4LHG2.cjs.map
+//# sourceMappingURL=agent-OKP4LHG2.cjs.map

package/dist/{agent-TLRURB7W.cjs.map → agent-OKP4LHG2.cjs.map}

@@ -1 +1 @@
-{"version":3,"sources":[],"names":[],"mappings":"","file":"agent-TLRURB7W.cjs"}
+{"version":3,"sources":[],"names":[],"mappings":"","file":"agent-OKP4LHG2.cjs"}