chore(deps): update ⬆️ mise-packages

[renovate]

This PR contains the following updates:

Package	Update	Change	Pending	Age	Adoption	Passing	Confidence
aqua:anthropics/claude-code	patch	2.1.218 → 2.1.220
aqua:earendil-works/pi	patch	0.82.0 → 0.82.1	0.83.0
aqua:sst/opencode	patch	1.18.4 → 1.18.5	1.18.9 (+3)
github:max-sixty/worktrunk	patch	v0.69.0 → v0.69.2	v0.70.0
github:modem-dev/hunk	patch	v0.17.4 → v0.17.6	v0.17.7
npm:@fsouza/prettierd	minor	0.28.0 → 0.29.0
usage	minor	3.5.6 → 3.6.0

Release notes are maintained in a PR comment by the renovate-release-notes-comment workflow.

---

Configuration

📅 Schedule: (in timezone America/Los_Angeles)

• Branch creation
  • Between 03:00 AM and 05:59 AM (* 3-5 * * *)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Enabled.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.

[scottames-github-bot]

Renovate Release Notes

Generated from Renovate's update table by the renovate-release-notes-comment workflow.

Packages that cannot be summarized from GitHub releases are listed explicitly below.

anthropics/claude-code (aqua:anthropics/claude-code)

v2.1.220: v2.1.220

Compare Source

What's changed

• Bug fixes and reliability improvements

v2.1.219: v2.1.219

Compare Source

What's changed

• Added Claude Opus 5 (claude-opus-5), now the default Opus model — 1M context, fast mode at $10/$50 per Mtok
• Added sandbox.network.strictAllowlist setting to deny non-allowlisted hosts for sandboxed commands without prompting
• Added DirectoryAdded hook that fires after /add-dir or the SDK register_repo_root control request registers a new working directory mid-session
• Added mcp_server_errors to the headless stream-json init event, listing --mcp-config entries skipped by config validation; terminal runs print a startup warning
• Added the workflowSizeGuideline settings key so the advisory Dynamic workflow size guideline can be set from any settings file; the /config row is hidden while one does
• Added nested subagent forwarding in stream-json: subagents spawned at depth-2+ now appear when --forward-subagent-text is set, keyed by their spawning Agent tool_use id
• Fixed claude -p text output dropping the answer already produced when a turn dies on a mid-stream API error
• Added HTTP status and error text to claude mcp list and /mcp when a server fails to connect, and a warning for MCP config values with hidden leading or trailing whitespace
• Fixed a permission you approved while a self-hosted runner was restarting being dropped when the session resumed, so the approved action now runs
• Fixed the Fable model row showing "Requires usage credits" for plans that include it, when a stale cache had baked the label in
• Fixed a SIGTERM arriving while a self-hosted runner was starting up leaving a stale active row until the lease expired; it now deregisters cleanly
• Added structured failure categories to self-hosted runner spawn and session failures, so hook errors, runner crashes and config errors can be told apart
• Fixed the /model picker showing the merged Opus row as plain "Opus" instead of "Opus (1M context)"
• Fixed copy-on-select inside GNU screen printing base64 into the terminal instead of copying the selection
• Fixed Remote Control clients keeping a stale fast-mode status after a model switch, reconnect, or failed org check
• Fixed CLAUDE_CODE_GIT_BASH_PATH on Windows exiting or being used as bash when the path isn't a bash/sh binary; it's now ignored with a warning
• Fixed Vim mode: pressing ← on an empty prompt now returns to the agent view from NORMAL mode, not just INSERT
• Fixed screen-reader mode rewriting the entire input line on every keystroke instead of echoing only the typed character
• Improved the "Remote Control is only available via api.anthropic.com" error to name the specific setting that caused it
• Improved claude --teleport to show which repo your current checkout points at when it doesn't match the session's repo
• Changed dynamic workflows to default to a medium size guideline (aim for fewer than 15 agents); pick another size or unrestricted with Dynamic workflow size in /config
• Changed managed MCP allowlist/denylist ${VAR} entries to resolve from the startup environment and managed-settings env instead of settings-file env
• Changed the /model picker to highlight only the newest model's name, so the highlight marks the new release rather than an arbitrary subset of the list
• Added the current default workflow size to the running-workflow status line, with a pointer to /config for changing it
• Removed Opus 4.7 from fast mode; /fast now applies to Opus 5 and Opus 4.8
• Updated the claude-api skill to default to Claude Opus 5, with a migration path from Opus 4.8
• Subagents can now spawn nested subagents up to depth 3 by default (was 1); set CLAUDE_CODE_MAX_SUBAGENT_SPAWN_DEPTH=1 to disable nesting

earendil-works/pi (aqua:earendil-works/pi)

v0.82.1: v0.82.1

Compare Source

New Features

• Claude Opus 5 — Available on Anthropic and Amazon Bedrock with adaptive thinking (including xhigh), inference profiles, and prompt caching. See Providers.
• Anthropic gateway bearer auth — ANTHROPIC_AUTH_TOKEN authenticates against Anthropic-compatible gateways that require Authorization: Bearer, including compaction and branch summaries. See Environment Variables or Auth File.
• Faster, more resilient model catalogs — pi.dev catalogs revalidate with If-None-Match so unchanged providers answer with an empty 304, and llama.cpp models stay listed across restarts. See llama.cpp.

Added

• Exposed the outputPad setting to custom message renderers. See Extensions (#7045 by @​xl0).
• Added inherited ANTHROPIC_AUTH_TOKEN bearer authentication for Anthropic-compatible gateways. See Providers (#5871).
• Added inherited Claude Opus 5 support for Anthropic and Amazon Bedrock with adaptive thinking, inference profiles, prompt caching, and preserved AWS validation messages (#7081 by @​unexge, #7083 by @​davidbrai).

Changed

• Changed pi.dev model catalog refreshes to revalidate with If-None-Match, so unchanged provider catalogs answer with an empty 304 instead of a full download.
• Changed inherited Radius OAuth device authorization, token exchange, and refresh requests to use the configured gateway directly.
• Changed inherited model loading errors to append the underlying cause, so auth failures such as OAuth refresh failed for openai-codex report the provider response instead of a bare wrapper message.

Fixed

• Fixed compaction and branch summaries for providers whose authentication resolves entirely to request headers (#5871)
• Fixed unavailable scoped models being hidden from /models, allowing them to be removed without editing settings manually (#6949, #7032 by @​christianklotz).
• Fixed startup context file discovery to skip directories that match context file names such as AGENTS.md, which produced EISDIR warnings (#7106 by @​mrexodia).
• Fixed the llama.cpp extension to persist its model catalog, so llama.cpp models stay listed before the first successful refresh. See llama.cpp (#7072 by @​davidbrai).

sst/opencode (aqua:sst/opencode)

v1.18.5: v1.18.5

Compare Source

Core

Bugfixes

• Improve Claude adaptive thinking handling across more response shapes.
• Avoid OpenAI Responses phase handling that could break some conversations.
• Preserve grep symlink paths in search results. (@​remixz)
• Preserve Mistral reasoning history across turns.
• Stabilize Mistral prompt caching.
• Use the correct prompt cache keys for each SDK.
• Fix MiniMax M3 thinking variant selection.

Desktop

Improvements

• Support the current server terminal transport.
• Support current-server review data in the desktop app.
• Update server discovery flows for current servers.
• Support current-server session actions, including prompts and commands.
• Render current-server session timelines.
• Stream current-server events in the desktop app.
• Detect legacy and current servers so the desktop app can work with both.

Bugfixes

• Restore optimistic timeline updates while responses are still streaming.
• Hide legacy-only features when connected to current servers.
• Preserve inline file mentions when sending prompts to legacy servers.
• Stop auto-accepting config permissions on current servers.
• Keep current servers out of the legacy layout.
• Show plain file contents in file-specific tabs instead of review diffs.
• Keep the prompt input agent toggle in sync.
• Keep paginated session timelines in the right order.
• Restore directory-scoped session status for legacy servers.
• Reload legacy session progress after hydration.

Thank you to 2 community contributors:

• @​dleopold:
  • fix(app): classify existing web profiles for layout transition (#38117)
• @​remixz:
  • fix(opencode): preserve grep symlink paths (#38581)

max-sixty/worktrunk (github:max-sixty/worktrunk)

v0.69.2: 0.69.2

Compare Source

Release Notes

Improved

• wt remove resolves every fsmonitor daemon in one lsof call: The end-of-command sweep forked one lsof per git fsmonitor--daemon on the machine — with core.fsmonitor enabled globally that is one daemon per repo ever touched, routinely over a hundred — and the spawn cost compounds under load rather than staying fixed per call. All daemons now resolve in a single call: 108 spawns to 1 on a live machine. (#3581)

Fixed

• Shell integration no longer deletes user data that merely quotes the init command: Three places decided whether text was worktrunk's by testing a blob for substrings, and two of them deleted on a wrong guess. wt config shell uninstall removed an rc line like alias setup='echo "run: wt config shell init fish | source"'; wt config shell install's legacy cleanup deleted a user's own ~/.config/fish/conf.d/wt.fish outright when the file happened to mention the init command. Both now require the command name in command position, not merely somewhere in the text. (#3589)

• Forge detection matches host labels, not substrings: wt switch pr:<n> and friends picked their provider with host.contains("github") / contains("dev.azure.com"), so a host that only spells the name inside a longer one — github-mirror.example, dev.azure.com.attacker.example — resolved as that provider. Matching is now label-wise, accepting the domain and its subdomains only; a self-hosted deployment this misses can still name its provider explicitly. (#3589)

• A crash mid-write no longer empties the file being written: wt config shell uninstall ended its rc rewrite by truncating in place, so a crash, a full disk, or a lost power cable between the truncate and the write left ~/.bashrc, ~/.zshrc, or a PowerShell profile empty or half-written, taking every line the user had ever added. Every user-file write now goes through one writer that writes a temp file beside the target and renames it into place. (#3585, #3591)

• wt step push refuses to run out of a half-finished operation: Mid-rebase the detached HEAD looks like a linear extension of the target, so wt step push main reported ✓ Pushed to main (1 commit) while moving the target branch onto a half-replayed history and leaving the rebase open; it now runs the same operation gate as wt step rebase and wt merge. The same change fixes an annotated-tag target always reporting as needing a rebase — the tag object's SHA was compared against a merge-base that peels it — and makes both push paths refuse a target worktree whose directory is gone, rather than --no-ff moving the ref over the stale registration. (#3578)

• Conflict markers can't reach a commit, and the refusal names the command you ran: wt step relocate --commit staged with git add -A and committed straight through an unresolved merge, reaching neither gate added for the other staging commands; every staging path is now gated on an unmerged index. wt merge already refused, but in a sub-step's name. (#3588, #3587)

• OpenCode activity markers land in the worktree they belong to: The plugin issued its marker commands through the process-global Bun shell without scoping them, so each ran in whatever the process-wide cwd happened to be — under concurrent parallel-agent sessions, a marker write could land in another session's worktree. (#3554, thanks @​4i3n6)

Documentation

• wt step rebase and wt step push render on the docs site: the only two of twelve step operations whose help was terminal-only. Both bodies are rewritten, correcting (among others) the claim that conflicts abort immediately — nothing aborts; the worktree is left mid-rebase with git's markers — and a wt step squash note promising a backup ref unconditionally, when a clean-tree squash writes none. (#3578)

• Troubleshooting no longer suggests disabling core.fsmonitor globally: the guidance for a wedged daemon keeps to the targeted fixes — kill the daemon serving that worktree, or let the next wt list respawn the live ones. (#3581)

Internal

• The Windows zip ships the signed binary: SignPath names its download after the GitHub artifact, so the signed zip landed beside the unsigned build as worktrunk-x86_64-pc-windows-msvc.zip.zip while the checksum step and the release upload both kept reading the original — so v0.69.1 shipped an unsigned binary under a green run and a completed signing request. (v0.69.0's unsigned binary was the separate upload bug fixed in #3566, where the request failed outright.) The signed file now replaces the built one only after it verifies, and a final step reads the shipped zip and reports any executable without a certificate. Signing remains non-blocking on a self-signed test certificate pending SignPath's OSS review, so Windows will not show a trusted publisher yet. (#3590)

Install worktrunk 0.69.2

Install prebuilt binaries via shell script

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/max-sixty/worktrunk/releases/download/v0.69.2/worktrunk-installer.sh | sh && wt config shell install

Install prebuilt binaries via powershell script

powershell -ExecutionPolicy Bypass -c "irm https://github.com/max-sixty/worktrunk/releases/download/v0.69.2/worktrunk-installer.ps1 | iex"; git-wt config shell install

Install prebuilt binaries via Homebrew

brew install worktrunk && wt config shell install

Download worktrunk 0.69.2

File	Platform	Checksum
worktrunk-aarch64-apple-darwin.tar.xz	Apple Silicon macOS	checksum
worktrunk-x86_64-apple-darwin.tar.xz	Intel macOS	checksum
worktrunk-x86_64-pc-windows-msvc.zip	x64 Windows	checksum
worktrunk-aarch64-unknown-linux-musl.tar.xz	ARM64 MUSL Linux	checksum
worktrunk-x86_64-unknown-linux-musl.tar.xz	x64 MUSL Linux	checksum

Install via Cargo

cargo install worktrunk && wt config shell install

Install via Winget (Windows)

winget install max-sixty.worktrunk && git-wt config shell install

Install via AUR (Arch Linux)

paru worktrunk-bin && wt config shell install

v0.69.1: 0.69.1

Compare Source

Release Notes

Improved

• wt switch statusline dims the dev-server URL until its port answers: The Claude Code statusline's dev-server URL now dims until something answers on its port, matching the wt list cell it already copied in every other respect. (#3561)

Fixed

• wt step rebase and wt merge refuse to run mid-operation: With a git operation already in progress (a conflicted rebase stop, or a killed git), wt step rebase <target> reported Already up to date and exited 0 over a conflicted, mid-replay tree — it asked "already rebased?" before consulting the worktree's operation state — and wt merge failed with a detached-HEAD error whose suggested git switch would have discarded the in-progress rebase. Both commit-replaying commands now detect an open operation up front and refuse with a clear message. (#3558)

• Shell-integration install/uninstall correctness: wt config shell now validates the integration command name and rejects malformed values (empty, leading -, or shell-unsafe characters) with a clear error rather than writing a broken rc line; recognizes manually-added or older-form integration lines on install, reporting already-configured instead of appending a duplicate; and, on uninstall, scans for worktrunk-managed wrapper files and rc lines by content marker, so integration installed under an alternate binary name (git-wt, …) is cleaned up regardless of the name it was installed under — while a user's own file that merely mentions wt config shell init is left untouched. (#2864)

• wt switch picker reflects a mid-session removal after a deleted-CWD recovery: When the picker recovered from a deleted working directory, accepting a row reused the startup-time repository snapshot — so an in-picker alt-x removal of a worktree or branch during that recovered session wasn't observed on accept. The accept path now rebuilds the repository. (#3557)

Internal

• Windows code-signing upload no longer double-zips: The unsigned Windows artifact was uploaded wrapped in an artifact-storage zip around the already-zipped binary, so SignPath couldn't locate wt.exe inside it and the (non-blocking) signing request failed. The upload now sets archive: false, submitting the real zip to SignPath. (#3566)

• crates.io publishing via trusted publishing (OIDC): The release workflow mints a short-lived crates.io credential per run via rust-lang/crates-io-auth-action instead of a stored CARGO_REGISTRY_TOKEN. (#3564)

Install worktrunk 0.69.1

Install prebuilt binaries via shell script

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/max-sixty/worktrunk/releases/download/v0.69.1/worktrunk-installer.sh | sh && wt config shell install

Install prebuilt binaries via powershell script

powershell -ExecutionPolicy Bypass -c "irm https://github.com/max-sixty/worktrunk/releases/download/v0.69.1/worktrunk-installer.ps1 | iex"; git-wt config shell install

Install prebuilt binaries via Homebrew

brew install worktrunk && wt config shell install

Download worktrunk 0.69.1

File	Platform	Checksum
worktrunk-aarch64-apple-darwin.tar.xz	Apple Silicon macOS	checksum
worktrunk-x86_64-apple-darwin.tar.xz	Intel macOS	checksum
worktrunk-x86_64-pc-windows-msvc.zip	x64 Windows	checksum
worktrunk-aarch64-unknown-linux-musl.tar.xz	ARM64 MUSL Linux	checksum
worktrunk-x86_64-unknown-linux-musl.tar.xz	x64 MUSL Linux	checksum

Install via Cargo

cargo install worktrunk && wt config shell install

Install via Winget (Windows)

winget install max-sixty.worktrunk && git-wt config shell install

Install via AUR (Arch Linux)

paru worktrunk-bin && wt config shell install

modem-dev/hunk (github:modem-dev/hunk)

v0.17.6: v0.17.6

Compare Source

What's Changed

• Avoid loading OpenTUI’s embedded native library for headless commands, preventing temporary native-library artifacts from leaking during help, version, session, daemon, markup, and non-interactive pager operations by @​benvinegar in fix(cli): lazy-load OpenTUI for headless commands modem-dev/hunk#590
• Optimize terminal cell-width measurement for faster rendering of diffs containing CJK, emoji, and other wide characters by @​kazu728 in perf(ui): optimize terminal cell width measurement (#579) modem-dev/hunk#586
• Republish the 0.17.5 application changes as 0.17.6 with downloadable binaries for every supported platform by @​benvinegar in chore(release): prepare 0.17.6 modem-dev/hunk#593

Full Changelog: modem-dev/hunk@v0.17.4...v0.17.6

fsouza/prettierd (npm:@​fsouza/prettierd)

v0.29.0: v0.29.0

Compare Source

Changed

• Minimum supported Node.js version is now 22 (dropped support for Node.js 20)

Updated dependencies

• prettier upgraded to ^3.9.5
• @​babel/parser upgraded to ^8.0.4
• @​typescript-eslint/typescript-estree upgraded to ^8.64.0

jdx/usage (usage)

v3.6.0: v3.6.0: effect= and four silent SpecCommand bugs

Compare Source

Adds an effect= annotation for classifying command side effects, and fixes four latent SpecCommand bugs uncovered while auditing the code paths a new field has to touch.

Added

• effect= on commands (#739 by @​jdx). Declare what running a command does to the world so docs, wrapper scripts, and AI agent allowlists can consume the same annotation instead of hand-maintaining their own lists:

  cmd "ls"        effect="read"        help="List installed tools"
  cmd "use"       effect="write"       help="Install a tool and add it to the config"
  cmd "uninstall" effect="destructive" help="Remove a tool"

  Effect	Meaning
  read	Only inspects state. Idempotent.
  write	Creates or modifies state, but removes nothing the user can't recreate.
  destructive	May delete or irreversibly overwrite. Deserves a confirmation prompt.

  Accepted as either a prop (effect="read") or a child node (effect "read"); unknown values are a parse error. The field is not inherited by subcommands (git remote and git remote remove do different things), and unset means unknown, not safe — consumers should treat missing values as "ask". Round-trips through KDL and renders an - **Effect**: line in generated markdown. Exported as usage::SpecCommandEffect.

Fixed

Four silent gaps in SpecCommand handling, all found by auditing every field against merge, the KDL serializer, and the docs model (#740 by @​jdx):

• deprecated was dropped by merge. An included spec could silently un-deprecate a command.
• examples were never written by the KDL serializer. spec.to_string() dropped every example.
• help_md / before_help_md / after_help_md were accepted only as props but serialized as child nodes. Any spec with markdown help failed to reparse with Error: unsupported cmd key help_md. They are now accepted as child nodes as well.
• restart_token never reached the docs model, so no template could render it.

Changed

• Missing SpecCommand fields are now a compile error (#740). merge and both From impls destructure their source with no .., so adding a field produces three compile errors pointing at exactly the places that owe it a decision. Runtime-derived fields are bound with a NOTE comment. A new round-trip test parses a spec exercising every field, serializes, reparses, and compares serde representations — catching the class of bug where the serializer writes something the parser rejects (which is how the help_md regression survived).

Full Changelog: jdx/usage@v3.5.7...v3.6.0

💚 Sponsor usage

usage is maintained by @​jdx, an open source developer for entire.io, the title sponsor of the jdx.dev open source tools including mise, aube, hk, and more. Work on usage is funded by sponsorships.

If usage powers CLI specs, docs, or completions for a tool you maintain or use, please consider sponsoring at jdx.dev. Every sponsorship helps the project stay independent and moving.

v3.5.7: v3.5.7: Mounted commands stop leaking mounting-CLI flags

Compare Source

A parser fix for shell completions inside mounted commands. Addresses jdx/mise#11282, where the mounting CLI's global flags leaked into mounted task completions and non-global flags before a task hid the task from the parser.

Fixed

• Mounted commands no longer inherit the mounting CLI's flags (#738). Three related defects in how the partial-parse's subcommand scan and re-parse interact with mount:

  1. Globals leaked into mounted commands. For a CLI like mise, where everything after a task name is forwarded to the task, mise run mytask --<TAB> used to offer --cd --env --jobs --locked --quiet --raw --silent --verbose --yes even though the mounted program rejects them (unexpected word: --silent). Worse, when a mounted task declared a flag whose name collided with a global (e.g. its own --env with choices), the global shadowed it and value completion fell back to file paths instead of the task's choices. Mounted commands are now marked internally and their own flags take precedence over inherited globals for their own names; a new ParseOutput::completion_flags() returns just the flags a completion should offer once a mount boundary has been crossed. Globals stay recognized before the mounted command, and mount scripts still receive them.

  2. A non-global flag hid the subcommand behind it. Phase 1 stopped scanning at the first non-global flag, so usage complete-word ... -- mise run --force build --bump '' failed with unexpected word: build. Known non-global flags are now consumed like globals and the scan continues; they are simply not forwarded to mounts. Unknown flags still stop the scan since their arity is unknown.

  3. Re-parsed prefix words could bind to the wrong flag. A global's value that appeared before a mounted command with a same-named flag would be re-validated against the mounted flag's choices and rejected. Phase 1 now records which flag each skipped word was read as, and Phase 2 respects that binding, so mycli --env prod run task --env <TAB> still parses --env prod as the global and offers the task's --env choices after the task name.

  Documentation for global flags and mounted commands has been added under docs/spec/reference/.

Full Changelog: jdx/usage@v3.5.6...v3.5.7

💚 Sponsor usage

usage is maintained by @​jdx, an open source developer for entire.io, the title sponsor of the jdx.dev open source tools including mise, aube, hk, and more. Work on usage is funded by sponsorships.

If usage powers CLI specs, docs, or completions for a tool you maintain or use, please consider sponsoring at jdx.dev. Every sponsorship helps the project stay independent and moving.