Skip to content

Release v0.8.0

Choose a tag to compare

View all tags
@github-actions github-actions released this 18 Jan 00:47
v0.8.0
bd637a0

Release v0.8.0

Installation

Install or upgrade LLxprt Code using npm:

npm install -g @vybestack/llxprt-code

Or use directly with npx:

npx @vybestack/llxprt-code

What's New

  • Added support for gemini-3-flash-preview model in OAuth mode

  • Added prompt_cache_key support to OpenAIResponsesProvider for improved caching behavior

  • Added interactive profile management UI with list, detail view, editor, and direct command access

  • Added interactive subagent management UI for delegating work to subagents

  • Added guided /profile create wizard for setting up new profiles

  • Added --continue flag for session resumption to resume interrupted conversations

  • Added welcome onboarding flow for first-time users

  • Added auto-enablement of ripgrep when available for faster search operations

  • Added timeout_ms parameter for grep tool to limit search duration

  • Added CLI configuration for task and shell tool timeout settings

  • Changed shell and task timeout from milliseconds to seconds for more intuitive configuration

  • Implemented compact todo widget with 25% height cap and collapse/expand functionality

  • Implemented sandwich compression to preserve conversation intent during summarization

  • Added markdown toggle (alt+@) to switch between rendered and raw markdown views

  • Added visual cue for closing dialogs with Escape key

  • Suppressed slash command execution and suggestions in shell output for cleaner interactions

  • Improved subagent output filtering based on profile emoji settings

  • Prevented trailing spaces on directory autocomplete

  • Fixed 403 OAuth token revoked errors with automatic refresh retry and bucket failover

  • Fixed 401 OAuth errors with clarified refresh retry behavior

  • Fixed OpenAI tool call ID preservation in pipeline mode to restore proper tool response matching

  • Fixed Claude Code OAuth compatibility for tokens starting with sk-ant-oat

  • Fixed OpenAI retry handling for timeouts

  • Fixed timeout settings autocomplete, profile support, and updated defaults

  • Fixed OpenAI tool output double-wrapping for error objects

  • Fixed Anthropic schema-aware parameter coercion for improved type error handling

  • Fixed tool call deduplication to prevent double execution

  • Fixed race condition in session restore by proactively initializing chat

  • Fixed token count race condition that caused footer flickering

  • Fixed queueing of tool completions during active stream processing

  • Fixed filtering of tool parts from retry requests to prevent duplicate tool_call blocks

  • Fixed race conditions in shell execution and tool scheduler

  • Fixed parallel tool execution hang with index-based publishing

  • Fixed scheduler hangs from cancelled tools not buffering results

  • Fixed tool continuation after cancel and handling of publishing errors

  • Fixed handling of 429/402/401 errors with immediate bucket failover trigger

  • Fixed OpenAI tool response formatting to return raw text without JSON stringification wrapper

  • Fixed todo_read/todo_write/todo_pause commands in noninteractive mode

  • Fixed prevent_eio crashes by adding error handlers to all setRawMode calls

  • Fixed auto-update to prevent ENOTEMPTY errors with atomic locking

  • Fixed memory refresh to properly load project-scoped memories in .llxprt/ directories

  • Fixed memory scope default from 'global' to 'project'

  • Fixed keytar credential storage by replacing with @napi-rs/keyring for cross-platform compatibility

  • Fixed greenscreen theming for about/auth/scrollbars to respect theme colors

  • Fixed colors for welcome onboarding text to respect theme colors

  • Fixed terminal mode management and cleanup on exit

  • Replaced /terminal-repair command with Ctrl+Alt+R hotkey for easier access

  • Fixed input handling in KeypressContext for improved block paste handling

  • Improved shell execution service reliability with buffered parallel tool execution

  • Fixed tool responses to require user approval for delete_line_range and insert_at_line in DEFAULT mode

Thanks

Huge thanks to the following contributors for their pull requests in this release:

All Changes

  • chore(release): v0.8.0 (bd637a0)
  • feat(gemini): add gemini-3-flash-preview to oauth models (#1169) (da45000)
  • fix: prevent OpenAI tool output double-wrapping for error objects (#1172) (ab47948)
  • Fix diff advisory in ui lockfiles (#1158) (a28d09b)
  • fix(deps): resolve 6 dependabot security alerts (#1154) (e51b99d)
  • fix(anthropic): add schema-aware parameter coercion for LLM type errors (#1149) (86cb6aa)
  • feat: Add prompt_cache_key support to OpenAIResponsesProvider (issue #1145) (#1147) (53b3fef)
  • fix(auto-update): prevent ENOTEMPTY errors with atomic locking (#1141) fixes #1137 (928704b)
  • fix: allow todo_write/todo_read/todo_pause in noninteractive mode (closes #1134) (#1140) (fcfe9a7)
  • fix: filter tool parts from retry requests to prevent duplicate tool_call blocks (#1136) (5ea41cd)
  • fix: queue tool completions that occur during active stream processing (#1127) (c3308ac)
  • fix: handle 403 OAuth token revoked errors with refresh retry and bucket failover (closes #1123) (#1125) (006f276)
  • fix(ci): skip E2E tests and coverage comment for docs-only PRs (closes #1114, closes #1119) (#1122) (72fa82a)
  • fix: eliminate token count race condition causing footer flickering (#1121) (3fcfeb2)
  • fix: make 'should combine multiple --allowed-tools flags' test platform-agnostic (#1120) (98ac1fc)
  • docs: fix Zed integration section and create zed-integration.md (#1118) (b86f903)
  • Revise Kimi subscription and usage instructions (a941a88)
  • docs: Main Documentation Gap Hitlist for 0.8.0 (#1112) (528812a)
  • feat: upstream sync to gemini-cli v0.12.0 (fixes #709) (#1088) (fd78fce)
  • fix(core): treat Anthropic overloaded_error as 429 for bucket failover (issue #1081) (#1103) (1d98351)
  • refactor(core): simplify nonInteractiveToolExecutor by removing duplicate code (issue #1057) (#1087) (cd34be5)
  • refactor(tests): replace snapshot tests with behavioral tests for GeminiMessage and ToolMessage (#1085) (83b9007)
  • fix: proactively initialize chat during session restore to prevent race condition (#1089) (bd60225)
  • Implement canonical tool ID pipeline (#1068) (48db4b0)
  • fix: replace keytar with @napi-rs/keyring for cross-platform credential storage (#1083) (0597ec3)
  • fix(core): consolidate multiple CoreToolScheduler instances into single shared instance (#1060) (106dd41)
  • fix(start): provide valid --localstorage-file path for DEV mode (#1079) (f8101cd)
  • [cli] feat: implement profile management UI with list, detail, editor, and direct command access (#1070) (6d9a01c)
  • Fix toolcall test nits and mouse buffer trim (#1074) (d9cf7fb)
  • Refactor sandbox env passthrough (#1073) fixed #1016 (759ecf1)
  • [cli] feat: add interactive subagent management UI (#1071) (dc5b6a7)
  • Clarify refresh retry behavior on 401 (#1072) closes #1015 (da2dad4)
  • feat: auto-enable ripgrep when available (#1069) fixes #960 (af7260f)
  • Fix greenscreen theming in about/auth/scrollbars (#1067) Fixes #697 Fixes #908 Fixes #910 (32849c4)
  • Fix memory refresh for new project memory (#1066) fixes #1065 (683129f)
  • Fix subagent delegation prompt gating (#1064) closes #1019 (032bc3d)
  • Delete CLAUDE.md (225c4ab)
  • Improve Anthropic prompt caching for history (#1062) closes #1046 (b540705)
  • Fix OpenAI retry handling for timeouts (#1061) fixes #1007 (2b15ec6)
  • fix: implement timeout settings autocomplete, profile support, and updated defaults (closes #1049) (#1058) (5910e53)
  • Fix Claude Code OAuth compatibility for tokens starting with sk-ant-oat (#1059) (3c76754)
  • fix(cli): deduplicate tool call requests to prevent double execution (issue #1044) (a051029)
  • Implement sandwich compression to preserve conversation intent (fixes #1011) (#1051) (afb6085)
  • [cli] feat: Add interactive /profile create wizard for guided profile setup (#1052) (cc4ef46)
  • feat: compact todo widget with 25% height cap and collapse/expand (#1043) closes #873 (8ca9831)
  • docs: remove duplicate header in README.md (07b2811)
  • chore: remove temporary files (BUCKET_FAILOVER_OPENAI_VERCEL.md, NOTES.md, issue1020.json, test-issue489-output.log) (6dcc178)
  • feat(subagent): apply emoji filtering to subagent output based on profile settings (#1041) (1e0b48f)
  • fix(debug): add factory pattern to DebugLogger to prevent memory leaks (#1039) fixes #228 (c074087)
  • fix(bucket-failover): check multiple config sources for failover handler (#1038) (5913179)
  • fix(oauth): wait indefinitely for bucket auth confirmation when prompt mode enabled (#1037) (67cdd3f)
  • feat: change shell and task timeout from milliseconds to seconds (#1035) (5e0767e)
  • Merge pull request #1032 from vybestack/20260104gmerge (b93ef8a)
  • style: run prettier on fileUtils.test.ts (ceef7cf)
  • Merge branch 'main' into 20260104gmerge (5ebf89d)
  • fix: correct test failures from incomplete cherry-picks (001aa55)
  • fix: enforce theme consistency for Ink Text components (#1033) (e740d6c)
  • fix: exclude GeminiMessage.test.tsx from vitest runs (1b39b2b)
  • fix: update GeminiMessage tests to mock RuntimeContext (4f023b0)
  • fix: restore Google LLC copyright in scripts/clean.js and scripts/deflake.js (1cf5a56)
  • fix: remove unused parseAllowedSubcommands function after merge (7ed33aa)
  • Merge branch 'main' into 20260104gmerge (5aa12c5)
  • fix(vscode): exclude tree-sitter from extension bundle (dd08b33)
  • style: apply prettier formatting (b251287)
  • docs(20260104gmerge): update Batch 13 status to IMPLEMENTED (8146173)
  • feat(core): add tree-sitter WASM shell parser with regex fallback (36e2696)
  • fix: add Colors.Foreground to all welcome onboarding text to respect theme colors (closes #1005) (#1030) (67d1c32)
  • docs(20260104gmerge): add tree-sitter implementation plan (f66ab30)
  • docs(20260104gmerge): add batch audit reports for all 49 batches (daee118)
  • fix: preserve OpenAI tool_call IDs in pipeline mode to restore tool response matching (fixes #981) (#1025) (2304bc7)
  • docs(20260104gmerge): batch 49 re-validation - loop detection skip (56d3f2f)
  • docs: fix Batch 48 notes with full validation outputs and AUDIT status (a36e49f)
  • docs(batch 48): re-validation - SKIP alternative architecture, functionally equivalent (d2ea714)
  • chore: remediate Batch 47 verification logging with full AGENTS.md command sequence (392aec2)
  • docs(batch 47): re-validation - SKIP incompatible architecture, superior LLxprt implementation (3b12e3e)
  • fix: change default memory scope from 'global' to 'project' (#1024) (45cdb53)
  • docs: remediate Batch 46 logging to NOTES.md-only requirement (c6fb708)
  • docs(batch 46): re-validation - FULLY VALIDATED - all 3 commits skipped or already implemented (36af84e)
  • docs: Add full npm run test output to Batch 45 NOTES.md section (e386dd4)
  • docs(batch 45): re-validation - FULLY VALIDATED - all commits skipped or already implemented (04fe41c)
  • docs: add full test/build outputs to Batch 44 NOTES.md (761fcbb)
  • docs(batch 44): re-validation - VERIFIED (superior DebugLogger implementation) (b603831)
  • docs: add explicit start command line to Batch 43 notes (cb561b5)
  • docs(batch 43): re-validation - SKIP incompatible architecture (6c9de57)
  • fix: prevent EIO crashes by adding error handlers to all setRawMode calls (#1023) (8a44afe)
  • docs(batch 42): add commit reference c6e26ef to documentation (655245b)
  • docs(batch 42): re-validation with full command output logs (c6e26ef)
  • docs(batch 41): re-validation with full AGENTS.md compliance - VERIFY/SKIP (7bc964e)
  • docs(batch 41): verify - SKIP/ALREADY_EXISTS - superior message bus and policy engine implemented (3c85df9)
  • docs: remediate Batch 40 validation with full AGENTS.md compliance (3f15011)
  • docs: merge Batch 40 notes from batch40-notes.md into NOTES.md (844e964)
  • docs(batch 40): verify/re-implement - wasm read test and deflake replace (012a8a6)
  • docs(batch 39): verify re-validation - static tool names already implemented on all tools (9af1642)
  • docs(merge): remediate Batch 38 NOTES.md with full validation outputs (4fa1714)
  • docs(batch 38): verify re-validation - all 3 commits already implemented with superior architecture (d29944d)
  • docs(batch 37): document SKIP/NO_OP re-validation for cc7e147 (b2403db)
  • docs(batch 36): fix AUDIT.md status for 995ae71 to SKIP/NO_OP (edaeccb)
  • docs: Remediate Batch 36 logging - fix AUDIT status and add full re-validation outputs (f34808f)
  • docs(batch 36): verify Batch 36 - NO_OP due to incompatible architecture (40f2937)
  • docs(batch 35): verify Batch 35 - 2 of 3 commits applied, all commands PASS (dc9ddfe)
  • docs: remediate batch 34 logging to NOTES.md (no separate notes file) (37154ce)
  • docs(batch 33): verify Batch 33 - already implemented or incompatible architecture (8d36ac7)
  • docs(batch 32): verify Batch 32 - abort signal retry support already implemented (NO_OP) (00c17c1)
  • docs: correct Batch 31 re-validation - all commands PASS (60827c9)
  • docs: mark Batch 31 as FAILED in audit and progress trackers (e721e6f)
  • docs: add Batch 31 re-validation with runtime command output logging (7f9e66c)
  • docs(batch 31): verify Batch 31 - tool name centralization already implemented with architectural improvements (ead314e)
  • docs(batch 30): remediate Batch 30 validation with full command outputs (d48ead3)
  • docs(batch 30): verify Batch 30 - UI hooks tests type errors resolved (6dd74c1)
  • docs(batch 29): complete Round 2 re-validation of Batch 29 (dac0f3a)
  • docs: remediate Batch 29 validation with full lint pass logs (e399ec3)
  • docs(batch 29): verify Batch 29 - already implemented with tool name centralization (5e6b77b)
  • feat: Add CLI configuration for task and shell tool timeout settings (#1021) (7c76cc5)
  • docs(batch 28): update re-validation section with corrected runtime command path (a3aa316)
  • docs: remediate Batch 28 logging with full validation outputs (4c2d8b9)
  • docs(batch 28): verify Batch 28 - already implemented as 4af9365 (925b616)
  • docs(batch 27): verify Batch 27 - NO_OP (already implemented with alternative type-safe approach) (0273c16)
  • docs(batch 26): verify Batch 26 - already implemented as 81be4bd with full validation (37133f9)
  • docs: remediate Batch 25 logging - add full validation outputs to NOTES.md (bce3850)
  • docs(batch 25): add Batch 25 verification summary (9338c99)
  • docs(batch 24): add full validation outputs to NOTES.md (91e2488)
  • docs(batch 23): add full validation outputs to NOTES.md, remove batch23-notes.md (5ac56bd)
  • docs: verify Batch 23 - skip (ui/components typecheck already enabled via architectural divergence) (1dc4b0c)
  • docs: add Batch 22 re-validation record (6119062)
  • docs: update PROGRESS.md and AUDIT.md with Batch 21 re-validation output (c21830c)
  • docs: Batch 21 re-validation - SKIP verified (LLxprt has superior DebugLogger) (2ae2ab5)
  • docs: batch 20 final validation - SKIP verified (4c1e806)
  • docs: Batch 19 full validation completed - VERIFIED (UPSTREAM-SKIP due to architectural divergence) (75745ef)
  • docs: Batch 18 full validation completed - VERIFIED (both commits SKIP'd) (0bf87ff)
  • docs: Batch 17 full validation completed - VERIFIED (all 5 commits already present) (fb1a1f2)
  • docs: add Batch 16 re-validation record to PROGRESS.md (2dd109f)
  • docs: batch 16 full validation completed - verified (3 PICKED, 2 SKIPPED) (1a87e34)
  • docs: add Batch 15 re-validation record with full command outputs (b3dd6eb)
  • docs: add Batch 14 re-validation record with full command outputs (fa3d27e)
  • docs: remediate Batch 13 verification failures with full outputs (ab84e8f)
  • docs: add Batch 13 re-validation record with full command outputs (26bcd12)
  • docs: add Batch 12 re-validation record with full command outputs (3c42427)
  • docs: add Batch 11 re-validation record with full command outputs (7eca03c)
  • docs: add Batch 10 re-validation record with full command outputs (8614f5f)
  • docs: add Batch 09 re-validation record with full command outputs (a3aba01)
  • docs: remediate Batch 08 verification logging (9f8cda1)
  • docs: add Batch 08 re-validation record with full command outputs (b6975ac)
  • docs: remediate Batch 07 re-validation with full command outputs (Deepthinker failure fix) (97befe8)
  • docs: add Batch 07 re-validation record with full command outputs (e274a11)
  • docs: add Batch 06 re-validation record with full command outputs (a593e28)
  • docs: add Batch 05 re-validation record with full command outputs (45920b7)
  • docs: add Batch 04 re-validation record with full command outputs (01a6e02)
  • docs: add Batch 03 re-validation record with full command outputs (d1bda9f)
  • docs: remediate Batch 02 verification - all required commands now PASS\n\n- npm run lint: PASS (exit code 0)\n- npm run typecheck: PASS (all 4 workspaces)\n- npm run build: PASS\n- node scripts/start.js --profile-load synthetic: PASS\n\nRoot cause was missing dist artifacts - build must run before lint/typecheck.\nFull command outputs logged in NOTES.md. (32d940e)
  • docs: add full command outputs to Batch 02 re-validation in NOTES.md (bc5e5c1)
  • docs: add Batch 02 re-validation record (2026-01-05) (4793d78)
  • docs: add Batch 01 re-validation record (2026-01-05) (be7366f)
  • fix(core): resolve conflict markers in fetch.ts (b8d8c18)
  • docs: ALL BATCHES COMPLETE (49/49) - v0.10.0 to v0.11.3 merge plan completed (7427af7)
  • fix(mcp): Include path in oauth resource parameter (#11654) (5b6901c)
  • fix(cli): fix flaky BaseSelectionList test (#11620) (a9ecf32)
  • Fix regression in handling shift+tab resulting in u in the input prompt. (#11634) (b1fc76d)
  • fix(scripts): add tsconfig linter to prevent adding files to the exclude list (#11602) (ba3c2f7)
  • fix(proxy): Add error handling to proxy agent creation (#11538) (f3d6f58)
  • docs: batches 36-40 complete - all marked SKIP/REIMPLEMENT/MANUAL_REVIEW due to conflicts (602a894)
  • docs: batches 31-35 complete - 2 PICKED, 7 items need REIMPLEMENT/MANUAL_REVIEW/SKIP (a992bc4)
  • fix(core): Preserve significant trailing spaces in gitignore patterns (#11536) (019f9da)
  • fix(core): address GCS path input (#11221) (fffbb87)
  • WIP: Uncommitted changes before Batch 31-35 (d21fdde)
  • feat(retry): abort retry loop on user abort signal (batch 32) (9bc69aa)
  • docs: batches 26-30 complete - 3 PICKED/REIMPLEMENTED, 2 SKIPPED (f2d0d25)
  • refactor(core): Add upstream tool name aliases for compatibility (fb8155a)
  • fix: Update web_fetch tool definition to instruct the model to provid… (#11252) (4af9365)
  • fix: Use shell:true default and add -I flag to system grep (81be4bd)
  • docs: update PROGRESS.md for batches 21-25 (all SKIPPED) (db555a3)
  • docs: batch 21-25 all SKIPPED - LLxprt architectural differences (343f9d5)
  • fix: trigger bucket failover immediately on 429/402/401 errors (#1006) (#1009) (ec58191)
  • docs: batch 20 complete - SKIPPED (LLxprt command tests diverged) (490a0ed)
  • docs: batch 19 complete - SKIPPED (Clearcut telemetry not in LLxprt) (df31fee)
  • docs: batch 18 complete - 2 SKIPPED (cosmetic + needs REIMPLEMENT) (c09cd50)
  • docs: batch 17 complete - 4 PICKED (a4dc52c, 32c4504, 5938d57, afe58d9), 1 SKIPPED (c039a8d)
  • fix: resolve typecheck errors from batch 17 cherry-picks (86af4ae)
  • fix: Unset selected auth type in integ test so that the local setting… (#11322) (afe58d9)
  • fix: make @file suggestions case-insensitive (#11394) (5938d57)
  • Only check for updates if disableUpdateNag is false (#11405) (32c4504)
  • fix(cli): enable and fix types for MCP command tests (#11385) (a4dc52c)
  • docs: batch 16 complete - 3 PICKED (a5ebead, f6d41e6, 8b6f764), 2 SKIPPED (f5edb0a)
  • feat: add missing visual cue for closing dialogs with Esc key (#11386) (8b6f764)
  • feat(cli): Suppress slash command execution and suggestions in shell … (#11380) (f6d41e6)
  • docs: update PROGRESS.md - batches 13-15 status corrected (2d3ad87)
  • fix: copy command delay in Linux handled (#6856) (a5ebead)
  • cherry-pick: upstream 3a1d376..3a1d376 batch 15 (8d48301)
  • docs: batch 14 complete - SKIPPED - functionality already exists (a88590a)
  • docs: batch 12 complete - SKIPPED - feature doesn't exist in LLxprt (3338361)
  • docs: batch 11 complete - SKIPPED due to different architecture (679d482)
  • docs: batch 10 complete - 3 commits all REIMPLEMENTED and COMMITTED (520c418)
  • fix(scripts): Improve deflake script isolation and unskip test (#11325) (upstream 991bd37) (bd104ab)
  • fix(scripts): Improve deflake script isolation and unskip test (#11325) (upstream 991bd37) (65b0e68)
  • fix: Add folder names in permissions dialog similar to the launch dialog (#11278) (upstream c71b749) (0e2efa6)
  • reimplement: refactor: Remove deprecated --all-files flag (#11228) (upstream 937c15c) (a35cb3d)
  • docs: batch 08 complete - 1x SKIP (already implemented), 1x PICK committed (fdfc122)
  • feat: Add markdown toggle (alt+m) to switch between rendered and raw markdown (81a4b03)
  • Use Node.js built-ins in scripts/clean.js instead of glob. (#11286) (c3d9e02)
  • docs: batch 07 complete - reimplemented Content-Type header check in fallback (cf33b18)
  • reimplement: fix(web-fetch): respect Content-Type header in fallback mechanism (#11284) (30a369b)
  • docs: batch 06 complete - 1x PICK committed, 2x SKIP already implemented (b486466)
  • feat: Do not add trailing space on directory autocomplete (#11227) (c527c3e)
  • docs: batch 03 (cfaa95a) complete - NO_OP already implemented (b2ebbe4)
  • cherry-pick: upstream 4f17eae..8c1656b batch 02 (f88b73f)
  • docs: batch 06 analysis - 1x PICK, 2x REIMPLEMENT (95d6c99)
  • docs: batch 05 (c9c633b) complete - tool names centralized (2e5f125)
  • reimplement: refactor: move web_fetch tool name to tool-names.ts (#11174) (upstream c9c633b) (19c6028)
  • docs: batch 03 (cfaa95a) analysis - REIMPLEMENT (3e5a9d7)
  • docs: add analysis notes to AUDIT.md preserving original decisions (50d4d7f)
  • docs: batch 03 (cfaa95a) and batch 04 (130f0a0) analysis complete (cc9fdd0)
  • reimplement: feat(core): wire up UI for ASK_USER policy decisions in message bus (upstream b8df8b2) (577de96)
  • feat: add --continue flag for session resumption (#999) (f9d76e1)
  • fix: format tool responses as raw text (no JSON stringify wrapper) (#1003) (85aea50)
  • fix(tools): require approval for delete_line_range and insert_at_line in DEFAULT mode (#1000) (42a8da1)
  • feat: Implement welcome onboarding flow for first-time users (#1001) (36815b5)
  • project-plans: reset 20260104gmerge batch logs (23154d8)
  • project-plans: add subagent execution runbook (a3c5726)
  • project-plans: seed 20260104gmerge plan folder (cc39f64)
  • fix: resolve child_process shell close without exit (#998) (c260738)
  • Merge pull request #994 from vybestack/issue982 (623910c)
  • fix(tests): address CodeRabbit feedback - use correct IContent property names (713ec88)
  • fix(compression): prevent empty compression split when history is tool-heavy (20c920b)
  • Merge pull request #992 from vybestack/issue966 (7a6066a)
  • fix(codex): use ONLY CODEX_SYSTEM_PROMPT in instructions field (2f91f47)
  • fix(codex): remove source_files disclosure from synthetic output (4f3b914)
  • fix(codex): always target AGENTS.md in synthetic call and include actual userMemory content (cba4cef)
  • fix(codex): stop injecting 'ignore system prompt' + pre-inject config file read (623fbcd)
  • Merge pull request #991 from vybestack/issue985 (c2b0fad)
  • fix: discover project-scoped memories in .llxprt/ directories during upward scan (3688db3)
  • Merge pull request #989 from vybestack/issue914 (8f179e2)
  • fix: add output truncation to PTY execution path (a726e45)
  • Merge pull request #988 from vybestack/issue983_987 (56af0c1)
  • fix: race conditions in shell execution and tool scheduler (95fe1c3)
  • Merge pull request #978 from vybestack/issue684 (405eb5e)
  • fix: pass resolved model name through logResponseStream call chain (af48f1c)
  • style: format code (eefe656)
  • fix: use resolved model name in /stats telemetry and fix LB tests (84e28d5)
  • fix: API stats tracking and theme violations in /stats model (d891fcb)
  • Merge pull request #976 from vybestack/issue975_974 (c6aca68)
  • fix: address CodeRabbit review comments (d405998)
  • fix: OAuth logout cache invalidation and ephemeral preservation (5f93680)
  • Merge pull request #973 from vybestack/issues971_962 (9e1b7f4)
  • feat(grep): add timeout_ms parameter for search operations (59bb897)
  • Merge pull request #972 from vybestack/issues959_916_847_199 (4fcf497)
  • fix: address exit hang and deduplicate ENABLE_MOUSE_EVENTS constant (6af6590)
  • fix: replace /terminal-repair command with Ctrl+Alt+R hotkey (c9b7087)
  • fix: terminal mode management and cleanup on exit (ac8e44f)
  • Merge pull request #970 from vybestack/issue968 closes #968 (62d78be)
  • refactor: DRY test helpers and add afterEach cleanup (3972f8e)
  • fix: prevent scheduler hangs from cancelled tools not buffering results (cf7a8f4)
  • fix: prevent tool continuation after cancel and handle publishing errors (a8e200a)
  • Merge pull request #965 from vybestack/issue956 (f35159d)
  • fix: improve block paste handling in KeypressContext (05990c3)
  • fix: always use passthrough mode for reliable bracketed paste (dbf6a25)
  • Merge pull request #961 from vybestack/issue945_948_951_952_957 (90da158)
  • fix: resolve parallel tool execution hang with index-based publishing (bb0519d)
  • fix: resolve cancel hang and unhandled rejection issues (c1cc7b6)
  • fix: address review feedback and add abort signal handling to grep (4ec9278)
  • fix: resolve tool scheduling, cancellation, and search issues (3651af2)
  • chore: update package-lock.json with version 0.8.0 (feb9158)
  • chore: bump version to 0.8.0 (aae75a4)
  • Merge pull request #947 from vybestack/issue943 (e347252)
  • fix(e2e): use positional prompt instead of stdin in sandbox tests (9ced98c)
  • fix(sandbox): fix stdin injection for boolean flags and empty image handling (1b6fbbb)
  • fix(sandbox): don't auto-enable sandbox in resolveSandboxEngine fallback (6a0710a)
  • fix(sandbox): require explicit opt-in for macOS seatbelt sandbox (bf1f25a)
  • fix(sandbox): avoid --prompt conflict when forwarding stdin to container (736927f)
  • test(integration): insert --profile-load after entrypoint (77942e3)
  • test(integration): fix profile-load argv insertion (cfc2e1f)
  • test(integration): clarify key/keyfile env requirement (7281d88)
  • test(integration): keep all prompt words last (db2abb0)
  • Merge pull request #954 from 5i2urom/feature/show-git-changes (e768184)
  • fix(core): add timeout to gitLineChanges and fix line count (80e8ee6)
  • feat(core): add showGitChanges to read_file/read_line_range (0579b8b)
  • test(integration): keep flags before positional prompt (e381cd8)
  • Merge pull request #949 from 5i2urom/feature/show-line-numbers (6b3e028)
  • fix(e2e): avoid --prompt conflict in sandbox:none (b022f34)
  • chore(core): format fixes (dc0b70d)
  • test(core): improve showLineNumbers coverage and fix docs (be60201)
  • test(e2e): skip google_web_search in CI (eb5700d)
  • test(e2e): avoid gemini-only web search (7e7489b)
  • feat(core): add showLineNumbers to read_file/read_line_range (c8eda13)
  • test(e2e): allow API key via keyfile (4cd5eb0)
  • chore(sandbox): tighten validation and messages (0b7380e)
  • fix(sandbox): create default profiles atomically (ef5c242)
  • chore(format): prettier (d56d409)
  • feat(sandbox): profiles, engine flag, ssh-agent passthrough (83fe819)
  • fix(sandbox): pause stdin before attach (e40c645)
  • Fix podman sandbox TTY allocation (d0d0680)
  • Remove sandbox visibility API step (0bb7846)
  • Retry sandbox visibility update (15e36a0)
  • Merge pull request #935 from vybestack/issue930 (66a83ea)
  • Fix sandbox visibility API package name (5d0a9cf)
  • Fix sandbox package visibility API path (4b94bfc)
  • Add Codex steering prompt in OAuth mode (9c2d7f2)
  • Add buildx setup for multi-arch sandbox builds (9d28e80)
  • Merge pull request #933 from vybestack/issue932 (c4b6be5)
  • Format release workflow (233304a)
  • Merge pull request #931 from vybestack/issue928_929 (3072bd1)
  • Fix sandbox visibility step (e3fb8c6)
  • Publish sandbox image for arm64 (82622e6)
  • Make sandbox image public in release workflow (58243d3)
  • Fix sandbox docs and add subagent guide (663c179)
  • Merge pull request #927 from vybestack/issue912_875 (a7eaa96)
  • feat: add mistral alias and gpt-5.2-codex (02dd638)
  • Merge pull request #906 from vybestack/issue904 (f351425)
  • Address CodeRabbit review suggestions (65e6257)
  • Merge pull request #921 from vybestack/issue919 (60dcdf7)
  • fix(auth): remove console.debug from OAuth providers (cb91196)
  • remove: delete SmartEdit tool (4d3a4e8)
  • Merge pull request #917 from vybestack/issue864 (193e706)
  • refactor(oauth): cache ProfileManager import and dedupe refresh token validation (ab034ef)
  • chore(core): use DebugLogger and simplify fetch cleanup (ffa185c)
  • Fix tool-call history for provider switching (issue #864) (2f3b902)
  • Merge branch 'main' into issue904 (d84d5d9)
  • Merge pull request #905 from vybestack/issue898 (8633652)
  • Merge branch 'main' into issue904 (462c230)
  • Merge branch 'main' into issue898 (9accf44)
  • Merge pull request #848 from vybestack/20251215gemerge (1da6681)
  • fix: address CodeRabbit review items (8095910)
  • format: prettier (2265c1e)
  • test(e2e): skip cyclic MCP test on Windows CI (5114629)
  • test(e2e): wait for MCP discovery in cyclic schema test (341647b)
  • test(core): stabilize settings remediation perf checks (0bf6121)
  • Merge remote-tracking branch 'origin/main' into 20251215gemerge (ecb083c)
  • OAuth: bucket-aware refresh + proactive renewals (7776281)
  • fix: prevent duplicate output for streaming tools in subagents (a026ec5)
  • fix(cli): propagate sandbox exit codes (aa509e8)
  • chore(integration): format test-helper (e73afcd)
  • chore(integration): fix lint (c5fbfaa)
  • test(integration): stabilize sandbox output and theme (b1e48e9)
  • fix(write_file): preserve trailing newline (7ffb402)
  • fix(replace): preserve trailing newline (5991e33)
  • Fix E2E: json output, tools allowlist, theme dialog (1481a0e)
  • perf(core): limit chardet buffer scan (fd6b755)
  • Fix Windows shell truncation test timeout (44e6b08)
  • chore: retrigger CI (5016cd4)
  • Fix PR 848 CI failures (443de70)
  • merge(main): resolve PR #848 conflicts (396ba03)
  • fix: address PR 848 review feedback (4ca0760)
  • plan: audit PR 848 review items (ff3f25f)
  • chore(lockfile): update workspace bin metadata (33c629e)
  • docs: add cherry-picking runbook (fd75cfe)
  • merge(main): resolve conflicts (b3388dc)
  • docs: drop unnecessary rerun guidance after format (c5d10ae)
  • docs: clarify tool-name aliases vs upstream divergence (a8208f0)
  • docs: adjust verification order for cherry-picking (ed2d6d9)
  • docs: remove marker-merge guidance from cherry-picking (38b6e6a)
  • fix(policy): allow non-interactive read tools by default (issue #707) (c90f48f)
  • test(core): restore literal emoji glyphs in emoji-filter tests (3a69efd)
  • test(core): avoid literal emoji characters in tests (a6ce2e4)
  • feat(core): unify non-interactive tool execution with scheduler (issue #707) (c7e0c00)
  • prep(core): shared tool governance + scheduler interactiveMode (issue #707) (b9fb189)
  • docs(plan): tighten unification cleanup + test scaffolding (issue #707) (5f0a173)
  • docs(plan): fix unification plan gaps (issue #707) (8f19228)
  • docs(plan): tool executor unification (issue #707) (3c79feb)
  • docs: clarify scheduler unification risks and call sites (24569a7)
  • test: add InputPrompt snapshot (621524c)
  • docs: preserve non-interactive gating in unification plan (d7fb5e2)
  • docs: make tool executor unification non-deferred (f104c31)
  • docs: update AUDIT.md - 907e51a (code-guide) now picked addresses #707 (04b2c63)
  • feat: add /code-guide slash command for codebase onboarding fixes #707 (c55bf13)
  • docs: normalize punctuation in tool executor plan (79ba0a4)
  • docs: correct tool executor unification plan (0f405fe)
  • docs: update AUDIT.md - cd91934 now picked addresses #707 (cd2fbab)
  • refactor: standardize env variable access to bracket notation fixes #707 (507de43)
  • docs: complete audit and add tool-executor unification plan addresses #707 (63d2da6)
  • docs: update AUDIT.md - Batch 51 (dabe161) now complete addresses #707 (5288d18)
  • docs: add batch audit for 20251215gemerge addresses #707 (9b8c52c)
  • feat: gate input until slash commands are loaded fixes #707 (5cb5c9f)
  • test: add MCP cyclic schema integration test addresses #27 (3096412)
  • reimplement: robust URL validation in web_fetch addresses #27 (753597b)
  • docs: document divergence from upstream replace test deletion addresses #27 (91a30ea)
  • reimplement: test harness yolo option + matchArgs addresses #27 (e8c441e)
  • test: add CI skip for ctrl-c interactive test addresses #27 (95431c9)
  • test: fix test mocks and add fetch retry test case addresses #27 (4d8aa07)
  • fix(e2e): use rmSync instead of rm -rf for e2e tests (4f067a2)
  • skip: cherry-pick 9e8c767 (record tool calls in non-interactive mode) (4755414)
  • fix: remove unused import and format (5f633db)
  • fix: Exit app on pressing esc on trust dialog at launch (#10668) (462e898)
  • Update system instructions for optimizing shell tool commands (#10651) (51c6473)
  • fix(compression): prevent unnecessary summarization when history is too short (#11082) (6810482)
  • perf(core): optimize Windows IDE process detection from O(N) to O(1) (#11048) (f1990f8)
  • feat(core): implement buffered parallel tool execution (9039042)
  • docs: document settings with showInDialog: true (8ae8ee4)
  • chore(settings): enable useSmartEdit by default (a90f574)
  • reimplement: rename expect methods (upstream a73b814) (4407297)
  • feat: add expectToolCallSuccess helper to TestRig (10df519)
  • refactor: use LLXPRT_CONFIG_DIR constant instead of hardcoded .gemini (9ab5761)
  • docs: add bundle command info to integration test docs (fa962dc)
  • Give explicit instructions for failure text in json-output.test.ts (#11029) (9bffe26)
  • fix(remove private) from gemini-cli-a2a-server (#11018) (5754041)
  • Deflake test. (#10932) (afb2fe6)
  • feat(tests): add InteractiveRun class for PTY test infrastructure (5c1d219)
  • feat(core): improve shell execution service reliability (#10607) (2011efb)
  • reimplement: update deployment.md -> installation.md (upstream 37678ac) (0a5bd47)
  • reimplement: simplify auth in interactive tests (upstream a64bb43) (fbbb014)
  • reimplement: fix rough edges around extension updates (upstream a6e00d9) (4cc35bf)
  • fix: update extension tests for LLxprt isWorkspaceTrusted signature (4805bed)
  • Show final install path in extension consent dialog and fix isWorkspaceTrusted check (#10830) (5759782)
  • reimplement: migrate web search tool to tool-names (upstream ab3804d) (cd439bd)
  • Show notification in screen reader mode (#10900) (c65fe0a)
  • fix: Usage of folder trust config flags in FileCommandLoader (#10837) (3ee0a7a)
  • docs: batch 25 execution record (NO-OP) addresses #707 (ce8ad41)
  • docs: batch 24 execution record (NO-OP) addresses #707 (7ffadef)
  • fix(test): flaky shell command test fix addresses #707 (abba675)
  • fix(test): Fix flaky shell command test using date command (#10863) (e719277)
  • fix(cli): memory refresh respects file filters addresses #707 (765495e)
  • fix(cli) : fixed bug #8310 where /memory refresh will create discrepancies with initial memory load ignoring settings/config for trusted folder and file filters (#10611) (1f0392b)
  • docs: batch 21 execution record (NO-OP) addresses #707 (a90e0dc)
  • feat(ui): add responsive margins and useFullWidth setting addresses #707 (a831b6d)
  • feat(a2a-server): add bin entry and refactor isMainModule addresses #707 (c272f45)
  • docs: batch 18 execution record + mock fix addresses #707 (74c6fdb)
  • clean up extension tests (#10857) (d7109b9)
  • fix some isWorkspaceTrusted mocks (#10836) (ae1ff54)
  • docs: batch 17 execution record addresses #707 (99c3128)
  • feat(core): Failed Response Retry via Extra Prompt (#10828) (a939e32)
  • docs: batch 16 execution record (NO-OP) addresses #707 (0cf4c39)
  • docs: batch 15 execution record addresses #707 (5388060)
  • test: skip flaky replace test (upstream 5e688b8) addresses #707 (6be2b11)
  • docs: batch 14 execution record addresses #707 (7e85c13)
  • fix(cli): prevent exit on non-fatal tool errors (#10671) (7350855)
  • docs: batch 13 execution record addresses #707 (0e521d6)
  • fix(tests): enable and update prompt for MCP add tool test (#10850) (b85747c)
  • Fix prompt to make it a bit more deterministic (#10848) (16fc60d)
  • fix(tests): log actual output in validateModelOutput on failure (#10843) (5474639)
  • docs: batch 12 execution record addresses #707 (2a66246)
  • fix(doc): update deflake command (upstream bd6bba8) addresses #707 (f88ca12)
  • docs: batch 11 execution record addresses #707 (4891eb0)
  • Fix hooks to avoid unnecessary re-renders (#10820) (bb8a5b7)
  • feat(ui): shorten context overflow message when <50% of limit (#10812) (d46080d)
  • docs: batch 10 execution record addresses #707 (b2d1136)
  • feat: support GOOGLE_CLOUD_PROJECT_ID fallback (upstream 0cd490a) addresses #707 (40c6f3a)
  • docs: batch 09 execution record addresses #707 (b8f7a32)
  • fix: Add a message about permissions command on startup in untrusted … (#10755) (671f8c4)
  • docs: batch 08 execution record addresses #707 (173ec6c)
  • feat(core): Stop context window overflow when sending chat (#10459) (90e5b98)
  • docs: batch 07 execution record (NO-OP) addresses #707 (bd979ad)
  • docs: batch 06 execution record addresses #707 (dee21d4)
  • fix(core): Use shell for spawn on Windows (#9995) (09c4fad)
  • docs: batch 05 execution record addresses #707 (6a992c6)
  • reimplement: add deflake docs section (upstream 8d8a2ab) addresses #707 (fd145ce)
  • docs: batch 04 execution record addresses #707 (c5ccb1f)
  • fix: populate argv.query/prompt from positional args and preserve JSON comments addresses #707 (eb32bbb)
  • citations documentation (#10742) (24d7d04)
  • fix(core): refresh file contents in smart edit given newer edits from user/external process (#10084) (8a9b759)
  • feat: prevent ansi codes in extension MCP Servers (#10748) (947de9c)
  • docs: batch 03 execution record addresses #707 (c048a71)
  • fix: post-cherry-pick fixes for Batch 03 addresses #707 (5970be2)
  • fix: ensure positional prompt arguments work with extensions flag (#10077) (dcf347e)
  • fix(mcp): fix MCP server removal not persisting to settings (#10098) (f8b3bb7)
  • chore(shell): Enable interactive shell by default (#10661) (90bc7aa)
  • Add script to deflake integration tests (#10666) (9728d9e)
  • fix(infra) - Remove auto update from integration tests (#10656) (82b61e2)
  • plan: add mandatory remediation subagent on verification failure addresses #707 (f828f0e)
  • docs: batch 02 verification record addresses #707 (5bae250)
  • fix: config.test.ts settings property paths addresses #707 (1bbdf78)
  • reimplement: bundle a2a-server (upstream 8ac2c68) addresses #707 (df79e75)
  • docs: batch 01 execution record (606aa6b)
  • Rationalize different Extension typings (#10435) (51cdc19)
  • plan: use typescript-master-coder (opus) for REIMPLEMENT batches (44db6c1)
  • plan: use actual subagent types for workflow (bf82c5a)
  • plan: add subagent workflow and verification requirements (839703d)
  • plan: add docs-only verification exception (675301f)
  • plan: add execution checklist and notes log (7af2cba)
  • plan: remove accidental compat shims (62b0122)
  • plan: modularize 20251215gemerge merge plan (b1cf697)
  • plan: 20251215 gemini-cli 0.9.0→0.10.0 merge plan (d7dd253)

Full Changelog: v0.7.0...v0.8.0

Contributors

file and 5i2urom
Assets 4
Loading