jeo-code 0.7.38 → 0.7.40

package/CHANGELOG.md

@@ -6,6 +6,26 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 The README mirrors the latest 5 entries — regenerate with `bun run changelog:sync`.
 
+## [0.7.40] - 2026-07-06
+_Fixes the release CI's standalone-binary build, broken since before v0.7.31 (last release where that job actually ran and passed) — the `release-binaries` job on v0.7.38 and v0.7.39 both failed silently on every platform target, so neither release has downloadable binaries attached._
+
+### Fixed
+- **`bun build --compile` failed on every target with `Could not resolve: "chromium-bidi/lib/cjs/...`.** `playwright-core`'s bundled `coreBundle.js` has a lazy `require("chromium-bidi/...")` reached only by its firefox/webkit-over-BiDi bridge (`connectBidiOverCdp`) — a code path jeo never takes (`src/agent/browser-session.ts` launches `chromium` over plain CDP only). `chromium-bidi` isn't installed (not a declared dependency of `playwright-core` itself), so Bun's compile-time bundler — which statically resolves every reachable `require()` regardless of whether it's ever called — failed the whole build. `scripts/ci-release-build-binaries.ts`'s `buildCommand` now passes `--external chromium-bidi`, leaving an unresolved (and, for jeo, unreachable) `require()` in the compiled binary instead of failing the build.
+
+### Verified
+- `bun run typecheck` clean; `bun test` 2376 pass / 0 fail.
+- `bun scripts/ci-release-build-binaries.ts --targets all` builds all 5 targets clean (darwin-arm64, darwin-x64, linux-x64, linux-arm64, win32-x64.exe) from a single macOS host, matching the CI runner's cross-compile setup; the darwin-arm64 output was smoke-run (`--version`) and the linux/windows outputs verified as valid ELF/PE binaries via `file`.
+
+## [0.7.39] - 2026-07-05
+_Fixes an unrecovered Anthropic refusal on Claude Fable 5: the new `stop_details.category: "reasoning_extraction"` refusal shape wasn't recognized, so the turn died with a raw, unfriendly error instead of engaging jeo's existing context-reset refusal-recovery ladder._
+
+### Fixed
+- **Anthropic's new `Refusal (reasoning_extraction): …` refusal message (Claude Fable 5) was not recognized as a refusal.** `isRefusalError` (`src/util/retry.ts`) only matched older shapes (`stop_reason=refusal`, `finish_reason=content_filter`, `(SAFETY)`, `(PROHIBITED_CONTENT)`, `(BLOCKLIST)`); Anthropic's `stop_details.category` refusal wording falls outside all of them, so `defaultRetryable` and `engine.ts`'s refusal ladder (free resend → context-reset + strip replayed thinking → strip `<project_context>` → abortable backoff) never engaged, and `friendlyProviderError` fell through to a raw passthrough — the user saw the bare provider text and the turn died. The regex now also matches the general `Refusal (<category>):` structural prefix so any current or future `stop_details.category` value is covered, not just `reasoning_extraction`. `friendlyProviderError` additionally appends a clarifying note specifically for `reasoning_extraction`: this category's wording reads like an accusation of extracting the model's internal reasoning, which is a common false positive for jeo's own thinking-block replay across multi-step tool use, not an actual violation.
+
+### Verified
+- `bun run typecheck` clean; full suite green.
+- New coverage in `test/refusal-recovery.test.ts`: `isRefusalError` recognizes the new shape (plus a synthetic future category, proving the fix isn't overfit to one literal string) while still matching every pre-existing shape and never false-positiving on an unrelated error; `friendlyProviderError` category-aware messaging; an end-to-end refusal-ladder engagement test via `runAgentLoop` with a mocked provider throwing the new shape.
+
 ## [0.7.38] - 2026-07-05
 _Ponytail review of 0.7.37's Telegram daemon: closes a real trust-boundary gap (any Telegram user could steer/cancel subagents, not just the paired chat) and a `getUpdates`-failure hot-loop, plus bloat/style cleanup — no new surface area, no config changes._
 

package/README.ja.md

@@ -207,11 +207,11 @@ 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`) -->
+- **[0.7.40]** (2026-07-06) — Fixes the release CI's standalone-binary build, broken since before v0.7.31 (last release where that job actually ran and passed) — the `release-binaries` job on v0.7.38 and v0.7.39 both failed silently on every platform target, so neither release has downloadable binaries attached.
+- **[0.7.39]** (2026-07-05) — Fixes an unrecovered Anthropic refusal on Claude Fable 5: the new `stop_details.category: "reasoning_extraction"` refusal shape wasn't recognized, so the turn died with a raw, unfriendly error instead of engaging jeo's existing context-reset refusal-recovery ladder.
 - **[0.7.38]** (2026-07-05) — Ponytail review of 0.7.37's Telegram daemon: closes a real trust-boundary gap (any Telegram user could steer/cancel subagents, not just the paired chat) and a `getUpdates`-failure hot-loop, plus bloat/style cleanup — no new surface area, no config changes.
 - **[0.7.37]** (2026-07-05) — Remote subagent visibility/control over Telegram (gjc Telegram-daemon parity, scoped to jeo's subagent surface only — no forum topics, no inline keyboards, no image attachments; see CHANGELOG 0.7.34 for what jeo intentionally does not replicate from gjc's full notification stack).
 - **[0.7.36]** (2026-07-04) — Removes two specific guardrails per explicit user direction: `jeo approve`'s human-only identity gate (the agent can now approve its own reviewed `jeo ralplan` plan) and the disk-staleness write/edit clobber guard (a write/edit no longer refuses just because the file changed on disk since the last read). Both removals were scoped narrowly after enumerating every guardrail in the codebase — the catastrophic-command hard block (`rm -rf /`, fork bombs, raw-disk `dd`/`mkfs`), the read-only subagent role tool gating, the deep-interview mutation lock, and the blind-edit (never-read-this-session) guard are all UNCHANGED.
-- **[0.7.35]** (2026-07-04) — `jeo team` now runs every plan step through the SAME subagent execution core as the `task`/`subagent` tools, and gains gjc-style concurrent workers: a plan can mark a contiguous run of independent steps to execute in parallel, each isolated in its own git worktree, merged back in order.
-- **[0.7.34]** (2026-07-04) — gajae-code 0.8.0/0.8.1 parity audit: closes a real cross-provider lone-surrogate JSON bug and makes turn-boundary compaction Esc-cancellable; the rest of gjc's 0.8.x release (Telegram daemon, ACP, cmux, coordinator MCP, tool-discovery-mode, an internal paged transcript viewport) is gjc-specific architecture jeo intentionally does not replicate, or was already-equivalent (confirmed by a 4-way parallel audit: ast_grep formatting, session-picker empty-filter clamping, PageUp/PageDown).
 
 See [CHANGELOG.md](CHANGELOG.md) for the full history.
 <!-- CHANGELOG:END -->

package/README.ko.md

@@ -207,11 +207,11 @@ 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`) -->
+- **[0.7.40]** (2026-07-06) — Fixes the release CI's standalone-binary build, broken since before v0.7.31 (last release where that job actually ran and passed) — the `release-binaries` job on v0.7.38 and v0.7.39 both failed silently on every platform target, so neither release has downloadable binaries attached.
+- **[0.7.39]** (2026-07-05) — Fixes an unrecovered Anthropic refusal on Claude Fable 5: the new `stop_details.category: "reasoning_extraction"` refusal shape wasn't recognized, so the turn died with a raw, unfriendly error instead of engaging jeo's existing context-reset refusal-recovery ladder.
 - **[0.7.38]** (2026-07-05) — Ponytail review of 0.7.37's Telegram daemon: closes a real trust-boundary gap (any Telegram user could steer/cancel subagents, not just the paired chat) and a `getUpdates`-failure hot-loop, plus bloat/style cleanup — no new surface area, no config changes.
 - **[0.7.37]** (2026-07-05) — Remote subagent visibility/control over Telegram (gjc Telegram-daemon parity, scoped to jeo's subagent surface only — no forum topics, no inline keyboards, no image attachments; see CHANGELOG 0.7.34 for what jeo intentionally does not replicate from gjc's full notification stack).
 - **[0.7.36]** (2026-07-04) — Removes two specific guardrails per explicit user direction: `jeo approve`'s human-only identity gate (the agent can now approve its own reviewed `jeo ralplan` plan) and the disk-staleness write/edit clobber guard (a write/edit no longer refuses just because the file changed on disk since the last read). Both removals were scoped narrowly after enumerating every guardrail in the codebase — the catastrophic-command hard block (`rm -rf /`, fork bombs, raw-disk `dd`/`mkfs`), the read-only subagent role tool gating, the deep-interview mutation lock, and the blind-edit (never-read-this-session) guard are all UNCHANGED.
-- **[0.7.35]** (2026-07-04) — `jeo team` now runs every plan step through the SAME subagent execution core as the `task`/`subagent` tools, and gains gjc-style concurrent workers: a plan can mark a contiguous run of independent steps to execute in parallel, each isolated in its own git worktree, merged back in order.
-- **[0.7.34]** (2026-07-04) — gajae-code 0.8.0/0.8.1 parity audit: closes a real cross-provider lone-surrogate JSON bug and makes turn-boundary compaction Esc-cancellable; the rest of gjc's 0.8.x release (Telegram daemon, ACP, cmux, coordinator MCP, tool-discovery-mode, an internal paged transcript viewport) is gjc-specific architecture jeo intentionally does not replicate, or was already-equivalent (confirmed by a 4-way parallel audit: ast_grep formatting, session-picker empty-filter clamping, PageUp/PageDown).
 
 See [CHANGELOG.md](CHANGELOG.md) for the full history.
 <!-- CHANGELOG:END -->

package/README.md

@@ -234,11 +234,11 @@ Huge thanks to [gajae-code](https://github.com/Yeachan-Heo/gajae-code) for the i
 ## Changelog
 
 <!-- CHANGELOG:START (auto-generated from CHANGELOG.md — run `bun run changelog:sync`) -->
+- **[0.7.40]** (2026-07-06) — Fixes the release CI's standalone-binary build, broken since before v0.7.31 (last release where that job actually ran and passed) — the `release-binaries` job on v0.7.38 and v0.7.39 both failed silently on every platform target, so neither release has downloadable binaries attached.
+- **[0.7.39]** (2026-07-05) — Fixes an unrecovered Anthropic refusal on Claude Fable 5: the new `stop_details.category: "reasoning_extraction"` refusal shape wasn't recognized, so the turn died with a raw, unfriendly error instead of engaging jeo's existing context-reset refusal-recovery ladder.
 - **[0.7.38]** (2026-07-05) — Ponytail review of 0.7.37's Telegram daemon: closes a real trust-boundary gap (any Telegram user could steer/cancel subagents, not just the paired chat) and a `getUpdates`-failure hot-loop, plus bloat/style cleanup — no new surface area, no config changes.
 - **[0.7.37]** (2026-07-05) — Remote subagent visibility/control over Telegram (gjc Telegram-daemon parity, scoped to jeo's subagent surface only — no forum topics, no inline keyboards, no image attachments; see CHANGELOG 0.7.34 for what jeo intentionally does not replicate from gjc's full notification stack).
 - **[0.7.36]** (2026-07-04) — Removes two specific guardrails per explicit user direction: `jeo approve`'s human-only identity gate (the agent can now approve its own reviewed `jeo ralplan` plan) and the disk-staleness write/edit clobber guard (a write/edit no longer refuses just because the file changed on disk since the last read). Both removals were scoped narrowly after enumerating every guardrail in the codebase — the catastrophic-command hard block (`rm -rf /`, fork bombs, raw-disk `dd`/`mkfs`), the read-only subagent role tool gating, the deep-interview mutation lock, and the blind-edit (never-read-this-session) guard are all UNCHANGED.
-- **[0.7.35]** (2026-07-04) — `jeo team` now runs every plan step through the SAME subagent execution core as the `task`/`subagent` tools, and gains gjc-style concurrent workers: a plan can mark a contiguous run of independent steps to execute in parallel, each isolated in its own git worktree, merged back in order.
-- **[0.7.34]** (2026-07-04) — gajae-code 0.8.0/0.8.1 parity audit: closes a real cross-provider lone-surrogate JSON bug and makes turn-boundary compaction Esc-cancellable; the rest of gjc's 0.8.x release (Telegram daemon, ACP, cmux, coordinator MCP, tool-discovery-mode, an internal paged transcript viewport) is gjc-specific architecture jeo intentionally does not replicate, or was already-equivalent (confirmed by a 4-way parallel audit: ast_grep formatting, session-picker empty-filter clamping, PageUp/PageDown).
 
 See [CHANGELOG.md](CHANGELOG.md) for the full history.
 <!-- CHANGELOG:END -->

package/README.zh.md

@@ -207,11 +207,11 @@ 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`) -->
+- **[0.7.40]** (2026-07-06) — Fixes the release CI's standalone-binary build, broken since before v0.7.31 (last release where that job actually ran and passed) — the `release-binaries` job on v0.7.38 and v0.7.39 both failed silently on every platform target, so neither release has downloadable binaries attached.
+- **[0.7.39]** (2026-07-05) — Fixes an unrecovered Anthropic refusal on Claude Fable 5: the new `stop_details.category: "reasoning_extraction"` refusal shape wasn't recognized, so the turn died with a raw, unfriendly error instead of engaging jeo's existing context-reset refusal-recovery ladder.
 - **[0.7.38]** (2026-07-05) — Ponytail review of 0.7.37's Telegram daemon: closes a real trust-boundary gap (any Telegram user could steer/cancel subagents, not just the paired chat) and a `getUpdates`-failure hot-loop, plus bloat/style cleanup — no new surface area, no config changes.
 - **[0.7.37]** (2026-07-05) — Remote subagent visibility/control over Telegram (gjc Telegram-daemon parity, scoped to jeo's subagent surface only — no forum topics, no inline keyboards, no image attachments; see CHANGELOG 0.7.34 for what jeo intentionally does not replicate from gjc's full notification stack).
 - **[0.7.36]** (2026-07-04) — Removes two specific guardrails per explicit user direction: `jeo approve`'s human-only identity gate (the agent can now approve its own reviewed `jeo ralplan` plan) and the disk-staleness write/edit clobber guard (a write/edit no longer refuses just because the file changed on disk since the last read). Both removals were scoped narrowly after enumerating every guardrail in the codebase — the catastrophic-command hard block (`rm -rf /`, fork bombs, raw-disk `dd`/`mkfs`), the read-only subagent role tool gating, the deep-interview mutation lock, and the blind-edit (never-read-this-session) guard are all UNCHANGED.
-- **[0.7.35]** (2026-07-04) — `jeo team` now runs every plan step through the SAME subagent execution core as the `task`/`subagent` tools, and gains gjc-style concurrent workers: a plan can mark a contiguous run of independent steps to execute in parallel, each isolated in its own git worktree, merged back in order.
-- **[0.7.34]** (2026-07-04) — gajae-code 0.8.0/0.8.1 parity audit: closes a real cross-provider lone-surrogate JSON bug and makes turn-boundary compaction Esc-cancellable; the rest of gjc's 0.8.x release (Telegram daemon, ACP, cmux, coordinator MCP, tool-discovery-mode, an internal paged transcript viewport) is gjc-specific architecture jeo intentionally does not replicate, or was already-equivalent (confirmed by a 4-way parallel audit: ast_grep formatting, session-picker empty-filter clamping, PageUp/PageDown).
 
 See [CHANGELOG.md](CHANGELOG.md) for the full history.
 <!-- CHANGELOG:END -->

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "jeo-code",
-  "version": "0.7.38",
+  "version": "0.7.40",
 
 
   "description": "Clean, highly optimized AI coding agent using spec-first loop",

package/src/util/provider-error.ts

@@ -61,7 +61,12 @@ export function friendlyProviderError(err: unknown): string {
     return `${provider} rejected the request: the conversation no longer fits the model's context window. Run /compact, drop large attachments, or start a fresh session.`;
   }
   if (isRefusalError(err)) {
-    return `${provider} declined to answer (safety refusal — no content returned). Usually a content classifier tripped on recently read file/search content: /retry, /compact or /new to drop the triggering context, or switch model with /model. If this persists on a Claude subscription (OAuth) login, Anthropic restricts third-party OAuth clients — set ANTHROPIC_API_KEY or use another provider.`;
+    const base = `${provider} declined to answer (safety refusal — no content returned). Usually a content classifier tripped on recently read file/search content: /retry, /compact or /new to drop the triggering context, or switch model with /model. If this persists on a Claude subscription (OAuth) login, Anthropic restricts third-party OAuth clients — set ANTHROPIC_API_KEY or use another provider.`;
+    const category = /Refusal \((\w+)\)/i.exec(msg)?.[1];
+    if (category === "reasoning_extraction") {
+      return `${base} (Category: reasoning_extraction — this classifier flags requests that look like probing the model's internal reasoning; jeo's own thinking-block replay across multi-step tool use is a common benign trigger, not an actual violation. /retry after a /compact usually clears it.)`;
+    }
+    return base;
   }
   if (status === 404) {
     return `${provider} does not recognize the requested model (HTTP 404). The id may be retired, gated, or mistyped — pick another with /model.`;

package/src/util/retry.ts

@@ -282,5 +282,5 @@ export function isUsageLimitError(err: unknown): boolean {
  *  capped backoff — gjc parity: a refusal is never terminal) is the correct recovery
  *  layer because it MUTATES the context between attempts. */
 export function isRefusalError(err: unknown): boolean {
-  return /stop_reason=refusal|finish_reason=content_filter|\(content_filter\)|\(SAFETY\)|\(PROHIBITED_CONTENT\)|\(BLOCKLIST\)/i.test(errorMessageOf(err));
+  return /stop_reason=refusal|finish_reason=content_filter|\(content_filter\)|\(SAFETY\)|\(PROHIBITED_CONTENT\)|\(BLOCKLIST\)|Refusal \(\w+\)/i.test(errorMessageOf(err));
 }