jeo-code 0.6.20 → 0.6.21

package/CHANGELOG.md

@@ -5,6 +5,15 @@ The format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 The README mirrors the latest 5 entries — regenerate with `bun run changelog:sync`.
++
+## [0.6.21] - 2026-06-18
+_Session thinking level now reaches the provider's actual reasoning depth, not just the token ceiling._
+
+### Fixed
+- **`/thinking`, `--thinking`, and `/fast` now change real provider reasoning depth.** Previously a live session thinking change only adjusted the per-step `maxTokens` budget; the provider's reasoning effort (Anthropic `thinking.budget_tokens`, OpenAI `reasoning_effort`, Gemini `thinkingConfig.thinkingBudget`) still came from the global `~/.jeo/config.json` `thinkingLevel`. `reasoningEffort` is now threaded from the session level through `AgentLoopOptions` → `ChatOptions` → the model manager, so the session setting actually controls how deeply the model reasons. When unset it still falls back to the global config.
+
+### Added
+- **`thinkingToReasoningEffort` mapping test** locking the session-level → provider-tier contract (minimal/low → low, medium → medium, high/xhigh → high, unset → undefined).
 
 ## [0.6.20] - 2026-06-18
 _Launch REPL internals decomposed into testable modules: `@mention` path completion, slash-command view renderers, and slash-command handlers extracted from the monolithic `launch.ts` into dedicated files with full unit-test coverage._

package/README.ja.md

@@ -200,11 +200,11 @@ CI は `.github/workflows/npm-publish.yml` で公開します — GitHub リリ
 ## 変更履歴 (Changelog)
 
 <!-- CHANGELOG:START (auto-generated from CHANGELOG.md — run `bun run changelog:sync`) -->
+- **[0.6.21]** (2026-06-18) — Session thinking level now reaches the provider's actual reasoning depth, not just the token ceiling.
 - **[0.6.20]** (2026-06-18) — Launch REPL internals decomposed into testable modules: `@mention` path completion, slash-command view renderers, and slash-command handlers extracted from the monolithic `launch.ts` into dedicated files with full unit-test coverage.
 - **[0.6.19]** (2026-06-18) — Post-turn hooks run once per batch (not per edit), local hook reads are mtime-cached, tool-result formatting is parallelized, and wrapped colored text keeps its tint.
 - **[0.6.18]** (2026-06-17) — Memory data-flow diagram and a README "Memory flow" section documenting the actual runtime behavior.
 - **[0.6.17]** (2026-06-17) — Legacy MEMORY.md migrates losslessly into the OKF concept bundle, with a one-shot command and a rollback toggle.
-- **[0.6.16]** (2026-06-17) — OKF memory grows a concept cross-link graph: 1-hop search expansion, bundle lint, graphify-optional.
 
 See [CHANGELOG.md](CHANGELOG.md) for the full history.
 <!-- CHANGELOG:END -->

package/README.ko.md

@@ -200,11 +200,11 @@ CI는 `.github/workflows/npm-publish.yml`로 배포합니다 — GitHub 릴리
 ## 변경 이력 (Changelog)
 
 <!-- CHANGELOG:START (auto-generated from CHANGELOG.md — run `bun run changelog:sync`) -->
+- **[0.6.21]** (2026-06-18) — Session thinking level now reaches the provider's actual reasoning depth, not just the token ceiling.
 - **[0.6.20]** (2026-06-18) — Launch REPL internals decomposed into testable modules: `@mention` path completion, slash-command view renderers, and slash-command handlers extracted from the monolithic `launch.ts` into dedicated files with full unit-test coverage.
 - **[0.6.19]** (2026-06-18) — Post-turn hooks run once per batch (not per edit), local hook reads are mtime-cached, tool-result formatting is parallelized, and wrapped colored text keeps its tint.
 - **[0.6.18]** (2026-06-17) — Memory data-flow diagram and a README "Memory flow" section documenting the actual runtime behavior.
 - **[0.6.17]** (2026-06-17) — Legacy MEMORY.md migrates losslessly into the OKF concept bundle, with a one-shot command and a rollback toggle.
-- **[0.6.16]** (2026-06-17) — OKF memory grows a concept cross-link graph: 1-hop search expansion, bundle lint, graphify-optional.
 
 See [CHANGELOG.md](CHANGELOG.md) for the full history.
 <!-- CHANGELOG:END -->

package/README.md

@@ -200,11 +200,11 @@ Required npm token permissions (repository secret `NPM_TOKEN`):
 ## Changelog
 
 <!-- CHANGELOG:START (auto-generated from CHANGELOG.md — run `bun run changelog:sync`) -->
+- **[0.6.21]** (2026-06-18) — Session thinking level now reaches the provider's actual reasoning depth, not just the token ceiling.
 - **[0.6.20]** (2026-06-18) — Launch REPL internals decomposed into testable modules: `@mention` path completion, slash-command view renderers, and slash-command handlers extracted from the monolithic `launch.ts` into dedicated files with full unit-test coverage.
 - **[0.6.19]** (2026-06-18) — Post-turn hooks run once per batch (not per edit), local hook reads are mtime-cached, tool-result formatting is parallelized, and wrapped colored text keeps its tint.
 - **[0.6.18]** (2026-06-17) — Memory data-flow diagram and a README "Memory flow" section documenting the actual runtime behavior.
 - **[0.6.17]** (2026-06-17) — Legacy MEMORY.md migrates losslessly into the OKF concept bundle, with a one-shot command and a rollback toggle.
-- **[0.6.16]** (2026-06-17) — OKF memory grows a concept cross-link graph: 1-hop search expansion, bundle lint, graphify-optional.
 
 See [CHANGELOG.md](CHANGELOG.md) for the full history.
 <!-- CHANGELOG:END -->

package/README.zh.md

@@ -200,11 +200,11 @@ CI 通过 `.github/workflows/npm-publish.yml` 发布 — GitHub 发布 release
 ## 更新日志 (Changelog)
 
 <!-- CHANGELOG:START (auto-generated from CHANGELOG.md — run `bun run changelog:sync`) -->
+- **[0.6.21]** (2026-06-18) — Session thinking level now reaches the provider's actual reasoning depth, not just the token ceiling.
 - **[0.6.20]** (2026-06-18) — Launch REPL internals decomposed into testable modules: `@mention` path completion, slash-command view renderers, and slash-command handlers extracted from the monolithic `launch.ts` into dedicated files with full unit-test coverage.
 - **[0.6.19]** (2026-06-18) — Post-turn hooks run once per batch (not per edit), local hook reads are mtime-cached, tool-result formatting is parallelized, and wrapped colored text keeps its tint.
 - **[0.6.18]** (2026-06-17) — Memory data-flow diagram and a README "Memory flow" section documenting the actual runtime behavior.
 - **[0.6.17]** (2026-06-17) — Legacy MEMORY.md migrates losslessly into the OKF concept bundle, with a one-shot command and a rollback toggle.
-- **[0.6.16]** (2026-06-17) — OKF memory grows a concept cross-link graph: 1-hop search expansion, bundle lint, graphify-optional.
 
 See [CHANGELOG.md](CHANGELOG.md) for the full history.
 <!-- CHANGELOG:END -->

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "jeo-code",
-  "version": "0.6.20",
+  "version": "0.6.21",
   "description": "Clean, highly optimized AI coding agent using spec-first loop",
   "type": "module",
   "main": "src/cli.ts",

package/src/agent/engine.ts

@@ -28,6 +28,7 @@ async function invokeCallLlm(history: Message[], options: {
   jsonMode: boolean;
   model?: string;
   maxTokens?: number;
+  reasoningEffort?: import("../ai/types").CallOptions["reasoningEffort"];
   signal?: AbortSignal;
   onUsage?: (u: { inputTokens?: number; outputTokens?: number }) => void;
   onRetry?: (attempt: number, err: unknown, delayMs: number) => void;
@@ -205,6 +206,11 @@ export interface AgentLoopOptions {
   model?: string;
   /** Max generation tokens per step (drives the thinking budget). */
   maxTokens?: number;
+  /** Provider reasoning depth (mapped from the live session thinking level). Threaded to
+   *  callLlm so `/thinking`, `--thinking`, and `/fast` reach the provider's actual reasoning
+   *  budget (Anthropic budget_tokens / OpenAI reasoning_effort / Gemini thinkingBudget), not
+   *  just the max-token ceiling. When unset the model-manager falls back to the global config. */
+  reasoningEffort?: import("../ai/types").CallOptions["reasoningEffort"];
   tools?: Record<string, ToolHandler>;
   signal?: AbortSignal;
   events?: AgentLoopEvents;
@@ -499,6 +505,7 @@ export async function runAgentLoop(history: Message[], opts: AgentLoopOptions):
               tools: nativeToolSchemasFor(Object.keys(tools)),
               model: opts.model,
               maxTokens: opts.maxTokens,
+              reasoningEffort: opts.reasoningEffort,
               signal: opts.signal,
               onUsage: u => { acc.inputTokens += u.inputTokens ?? 0; acc.outputTokens += u.outputTokens ?? 0; sawUsage = true; },
               onToken,

package/src/agent/loop.ts

@@ -10,6 +10,11 @@ export interface ChatOptions {
   systemPrompt?: string;
   temperature?: number;
   maxTokens?: number;
+  /** Provider reasoning depth (mapped from the live session thinking level). When set it
+   *  overrides the global config-derived effort, so `/thinking` and `--thinking` reach the
+   *  provider's actual reasoning budget (Anthropic budget_tokens / OpenAI reasoning_effort /
+   *  Gemini thinkingBudget), not just the max-token ceiling. */
+  reasoningEffort?: import("../ai/types").CallOptions["reasoningEffort"];
   jsonMode?: boolean;
   signal?: AbortSignal;
   onUsage?: (usage: import("../ai/types").Usage) => void;

package/src/commands/launch.ts

@@ -36,7 +36,7 @@ import { callLlm, type Message } from "../agent/loop";
 import { friendlyProviderError } from "../util/provider-error";
 import { readGlobalConfig, saveConfigPatch } from "../agent/state";
 import { rememberModelPatch, recentModelsForDisplay } from "../agent/model-recency";
-import { describeModel, describeAllProviders, thinkingMaxTokens, discoverModels, flattenModels, resolveSelection, catalogMetadata, resolveRoleModel, CODEX_MODELS, qualifyModelId } from "../ai";
+import { describeModel, describeAllProviders, thinkingMaxTokens, thinkingToReasoningEffort, discoverModels, flattenModels, resolveSelection, catalogMetadata, resolveRoleModel, CODEX_MODELS, qualifyModelId } from "../ai";
 import type { ProviderModelsResult, PickEntry, ProviderName, ModelRole, ThinkLevel } from "../ai";
 import { readGoalState, writeGoalState, clearGoalState, verifyGoal } from "../agent/goal-verifier";
 
@@ -786,6 +786,7 @@ export async function runLaunchCommand(args: string[]): Promise<void> {
           maxSteps: flags.maxSteps,
           model: sessionModel,
           maxTokens: sessionThinking ? thinkingMaxTokens(sessionThinking) : undefined,
+          reasoningEffort: sessionThinking ? thinkingToReasoningEffort(sessionThinking) : undefined,
           signal: ac.signal,
           steer: drainSteer,
           events: wrapEvents(withStepPersistence({ ...withToolDetailCapture(tui ? tui.events() : streamEvents), onBeforeDone }, persistTurnTail), opik),
@@ -804,6 +805,7 @@ export async function runLaunchCommand(args: string[]): Promise<void> {
             budget: { maxExtensions: 0 },
             model: sessionModel,
             maxTokens: sessionThinking ? thinkingMaxTokens(sessionThinking) : undefined,
+            reasoningEffort: sessionThinking ? thinkingToReasoningEffort(sessionThinking) : undefined,
             signal: ac.signal,
             steer: drainSteer,
             events: wrapEvents(withToolDetailCapture(tui ? tui.events() : streamEvents), opik),