Releases: cbusillo/code
Release v0.6.106
@just-every/code v0.6.106
See CHANGELOG.md for details.
Release v0.6.105
@just-every/code v0.6.105
See CHANGELOG.md for details.
Release v0.6.104
@just-every/code v0.6.104
See CHANGELOG.md for details.
Release v0.6.103
@just-every/code v0.6.103
See CHANGELOG.md for details.
Release v0.6.102
@just-every/code v0.6.102
See CHANGELOG.md for details.
Release v0.6.101
@just-every/code v0.6.101
See CHANGELOG.md for details.
Release v0.6.100
@just-every/code v0.6.100
See CHANGELOG.md for details.
Code overlay-v0.6.98.1
What's Changed
This overlay release carries local fork updates on top of upstream Code 0.6.98.
It includes 233 listed changes since overlay-v0.6.97.8.
Fixes
- fix(protocol): refresh app-server schemas
- fix(core): backport upstream request compatibility
- fix: preserve exact turn diffs after partial apply_patch failures (openai#21518)
- fix(tui): clear first inline viewport render (openai#21450)
- fix(tui): skip terminal probes in detached tmux
- fix(tui): persist ctrl-c draft via app event (openai#21397)
- fix(tui): keep Ctrl-C stashed drafts after /clear (openai#21351)
- fix(bwrap): emit libcap after standalone archive (openai#21285)
- fix(linux-sandbox): isolate Linux sandbox synthetic mount registry per user for shared codex use case (openai#21234)
- fix(linux-sandbox): avoid panic on bwrap build failures (openai#21127)
- fix(sandboxing): Bound advisory system bwrap startup probe (openai#20111)
- fix(tui): external editor expansion for same-size large pastes (openai#21190)
- fix(tui): make /copy work inside tmux without passthrough (openai#20207)
- fix(tui): use shared paste burst interval on Windows (openai#18914)
- fix(tui): support modified backspace/delete keys (openai#21058)
- fix(linux-sandbox): fall back when system bwrap lacks perms (openai#20628)
- fix(tui): bound startup terminal probes (openai#20654)
- fix: reduce ConfigBuilder::build stack usage (openai#20650)
- fix(tui): restore alt-enter newline alias (openai#20535)
- fix: cargo deny (openai#20627)
Features
- feat(tui): add upstream-compatible slash commands
- feat: Update plugin share settings with discoverability (openai#21637)
- feat: enable AWS login credentials for Bedrock auth (openai#21623)
- feat(app-server, threadstore): Thread pagination APIs and ThreadStore contract (openai#21566)
- feat(local): add Every Code worker service entrypoint
- feat: Expose plugin share metadata in shareContext (openai#21495)
- feat: make built-in MCPs first-class runtime servers (openai#21356)
- feat(tui): attach Every Code origin to remote inbox sessions
- feat: Add marketplace source filtering and plugin share context (openai#21419)
- feat(local): add Every Code worker wrapper
- feat: move auto vaccum (openai#21378)
- feat(app-server): move v2
sessionIdontoThread(openai#21336) - feat: return session ID from thread/fork (openai#21332)
- feat: include thread ID in MCP turn metadata (openai#21329)
- feat: add
session_id(openai#20437) - feat: Add plugin share access controls (openai#21124)
- feat(scripts): add local Every Code issue poller
- feat(tui): redesign session picker (openai#20065)
- feat(tui): route /diff through workspace commands (openai#21001)
- feat(tui): add raw scrollback mode (openai#20819)
- feat: add normalized matching to memory search (openai#21205)
- feat: support windowed multi-query memory search (openai#21204)
- feat(tui): improve TUI keymap coverage (openai#20798)
- feat(tui): add PR summary statusline items (openai#20892)
- feat(tui): add keymap debug inspector (openai#20794)
- feat(app-server): always return limited thread history (openai#20682)
- feat: support multi-query memories search (openai#21004)
- feat: add context lines to memories MCP search (openai#20997)
- feat: paginate memories MCP search results (openai#20996)
- feat: make memories MCP list shallow (openai#20994)
- feat: paginate MCP memories list (openai#20993)
- feat: add max_lines to memories MCP read (openai#20991)
- feat: add line offsets to memory read MCP (openai#20986)
- feat: add remote compaction v2 Responses client path (openai#20773)
- feat: memories mcp v1 (openai#20622)
- feat: support template interpolation in multi-agent usage hints (openai#20973)
- feat: export and replay effective config locks (openai#20405)
- feat: seed ad-hoc memory extension instructions (openai#20606)
- feat: ad-hoc instructions (openai#20602)
Maintenance Updates
- docs(changelog): update for v0.6.98 [skip ci]
- chore(release): 0.6.98 [skip ci]
- chore: spawn MCP for memories (openai#21214)
- test: isolate app-server-client in-process test state (openai#21328)
- chore(app-server-protocol): split v2 API definitions into modules (openai#21251)
- chore: add minimal proxy egress diagnostics (openai#21220)
- memories-mcp: reject symlink traversal in local backend (openai#21010)
- Remove local docs and specs (openai#20896)
- chore: improve remember prompt (openai#20610)
- chore: allow memories edition (openai#20600)
CI and Release Infrastructure
- ci: trigger rusty-v8 releases from tags (openai#21259)
- ci: cross-compile Windows Bazel clippy (openai#20701)
- ci: cross-compile Windows Bazel tests (openai#20585)
Other Updates
- Refresh codex-rs mirror to upstream/main
- [codex] make shutdown pending-touch test deterministic (openai#21550)
- Omit service_tier from remote /responses/compact requests under API auth (openai#21676)
- Remove exec research preview banner wording (openai#21683)
- Fix feature request Contributing link (openai#21688)
- Fix issue template labels (openai#21686)
- Fix duplicate CLI issue template description (openai#21685)
- Remove skills list extra roots (openai#21485)
- [codex-analytics] plumb protocol-native review timing (openai#21434)
- Revert "Use
--lockedin cargo build and lint invocations" (openai#21646) - Send response.processed after remote compaction v2 (openai#21642)
- Add CODEX_HOME environments TOML provider (openai#20666)
- Route view_image through selected environments
- Make environment providers own default selection (openai#20665)
- [codex] Remove remote thread store implementation (openai#21596)
- Add stdio exec-server client transport (openai#20664)
- Use
--lockedin cargo build and lint invocations (openai#21602) - [codex] Apply a Dependabot cooldown of 7 days (openai#21599)
- codex-otel: add configurable trace metadata (openai#21556)
- Disable empty Cargo test targets (openai#21584)
- Ensure all mentions of cargo-install are --locked (openai#21592)
- [codex] Fully qualify hash-pins in GitHub Actions (openai#21436)
- Add a Cargo build profile for benchmarking (openai#21574)
- Use descriptive names for Cargo profile options (openai#21582)
- Grant sandbox users access to desktop runtime bin (openai#21564)
- Route ThreadManager rollout path reads through thread store (openai#21265)
- [codex] Fix pathless thread summaries (openai#21266)
- Move thread name edits to ThreadStore (openai#21264)
- Upgrade
cargo-shearto 1.11.2 (openai#21547) - [codex] Move tool specs onto handlers (openai#21461)
- app-server: refresh live threads from latest config snapshot (openai#21187)
- add top-level remote-control command (openai#21424)
- [codex] Remove string-keyed MCP tool maps (openai#21454)
- [codex-analytics] add tool review event schema (openai#18747)
- device-key: clean up unused crate (openai#21487)
- [codex] add account id to feedback uploads (openai#21498)
- Make turn diff tracking operation backed (openai#21180)
- Show plugin hooks in plugin details (openai#21447)
- [codex] fix PluginListParams test initializer (openai#21494)
- Revert state DB injection and agent graph store (openai#21481)
- [codex] Parallelize skills list cwd loading (openai#21441)
- [codex] allow shared config reads in app-server queue (openai#21340)
- [codex] Add OpenAI Developers to tool suggest allowlist (openai#21423)
- [codex] Delete tool handler plan indirection (openai#21427)
- Revert "Move skills watcher to app-server" (openai#21460)
- Add compact lifecycle hooks (started by vincentkoc - external contrib) (openai#19905)
- [codex] Move tool specs into core handlers (openai#21416)
- Move skills watcher to app-server (openai#21287)
- Document Codex git commit attribution config (openai#21379)
- [codex] Fix Windows sandbox git safe.directory for worktrees (openai#21409)
- [codex-analytics] emit tool item events from item lifecycle (openai#17090)
- [codex-tui] pass thread source for tui threads (openai#21401)
- [codex] Split tool handlers into separate files (openai#21395)
- [codex] Dedupe fallback model metadata warnings (openai#21090)
- Avoid hard-coded environment context shell (openai#21390)
- Avoid noisy OTEL diagnostics in codex exec (openai#21107)
- Route opted-in MCP elicitations through Guardian (openai#19431)
- [codex] Handle git pagination flags by position (openai#21381)
- Remove core MCP list tools op (openai#21281)
- vendor: update bubblewrap to 0.11.2 (openai#21389)
- [codex] Coordinate OpenAI docs sample with API key setup (openai#21263)
- rollout: coalesce thread updated_at touches (openai#21367)
- [codex] Add response.processed websocket request (openai#21284)
- Move message history out of core (openai#21278)
- 2- Use string service tiers in session protocol (openai#20971)
- [codex] fix builtin MCP Windows path test (openai#21350)
- Move installation ID resolution out of core startup (openai#21182)
- Propagate cache key and service tiers in compact (openai#21249)
- Revert "feat: support template interpolation in multi-agent usage hints" (openai#21337)
- Support Codex Apps auth elicitations (openai#19193)
- release: bundle bwrap with Linux codex DotSlash artifact (openai#21312)
- [mcp] Return Accept early per feedback. (openai#21277)
- Preserve session MCP config on refresh (openai#21055)
- app-server: align dynamic tool identifiers with Responses API (openai#20724)
- [codex-analytics] rework thread_source for thread analytics (openai#20949)
- Expose plugin manifest keywords in app server (openai#21271)
- [codex] Remove legacy ListSkills op (openai#21282)
- [codex] Remove unused ListModels op (openai#21276)
- release/npm: bundle standalone bwrap on Linux (openai#21257)
- Share Git safe-command logic on Windows (openai#21275)
- Add model and reasoning effort to MCP turn metadata (openai#21219)
- [codex] Move thread naming to app server (openai#21260)
- release: publish standalone bwrap artifacts (openai#21256)
- linux-sandbox: use standalone bundled bwrap (openai#21255)
- fix build (openai#21261)
- codex: use ThreadStore history for core review forks (openai#20577)
- Add cloud executor registration to exec-server (openai#19575)
- Inject state DB, agent graph store (openai#20689)
- Enable V8 sandboxing for source-built builds (openai#21146)
- [codex] fix TUI turn items view fixtures (openai#21243)
- Auto-deny MCP elicitations for Xcode 26.4 clients (openai#21113)
- [codex] Split tool handlers by tool name (openai#20687)
- add turn items view to app-server turns (openai#21063)
- [codex] Use shared app-server JSON-RPC error helpers (openai#21221)
- hook trust metadata and enforcement (openai#20321)
- Route process tools to selected environments (openai#20647)
- [codex-analytics] add tool item event schemas (openai#17089)
- app-server: ignore persist_extended_history param (openai#21225)
- revert legacy notify deprecation (openai#21152)
- [codex] Fix fork --last cwd filtering (openai#21089)
- [codex] Fix TUI large paste placeholder numbering after Ctrl+C ...
Code overlay-v0.6.97.8
What's Changed
This overlay release carries local fork updates on top of upstream Code 0.6.97.
It includes 1 listed changes since overlay-v0.6.97.7.
Fixes
- fix(tui/limits): refresh expired reset windows over HTTP
Verification
- Binary Release workflow completed successfully for Linux x64, macOS x86_64, and macOS aarch64.
- Release smoke tests ran for the Linux x64 and macOS aarch64 binaries before packaging.
- macOS x86_64 was built and packaged, but not smoke-tested in workflow.
- SHA-256 checksums are included in SHA256SUMS.txt.
Build
Built from overlay-v0.6.97.8 at 6586af0afeb0bb1bde369e7c8f509b3251f5754f.
Previous overlay release: overlay-v0.6.97.7.
Assets
code-aarch64-apple-darwin.tar.gzcode-aarch64-apple-darwin.zstcode-linux-x64.tar.gzcode-x86_64-apple-darwin.tar.gzcode-x86_64-apple-darwin.zstRELEASE_NOTES.txtSHA256SUMS.txt
Code overlay-v0.6.97.7
What's Changed
This overlay release carries local fork updates on top of upstream Code 0.6.97.
It includes 1 listed changes since overlay-v0.6.97.6.
Fixes
- fix(limits): preserve hint-only account blocks
Verification
- Binary Release workflow completed successfully for Linux x64, macOS x86_64, and macOS aarch64.
- Release smoke tests ran for the Linux x64 and macOS aarch64 binaries before packaging.
- macOS x86_64 was built and packaged, but not smoke-tested in workflow.
- SHA-256 checksums are included in SHA256SUMS.txt.
Build
Built from local/cbusillo-overlay at 4da04a7abc116aa23c490bb2e102e29803a6d4b4.
Previous overlay release: overlay-v0.6.97.6.
Assets
code-aarch64-apple-darwin.tar.gzcode-aarch64-apple-darwin.zstcode-linux-x64.tar.gzcode-x86_64-apple-darwin.tar.gzcode-x86_64-apple-darwin.zstRELEASE_NOTES.txtSHA256SUMS.txt