* feat: let operators compose full-code apps when the workspace grants it Adds operator_settings.builder_apps, the second builder right, granted independently of flows because the two are not equally verifiable: a composed flow is checked in full server-side, while a full-code app is a browser-built bundle no check can read. The app right rests instead on forced sandbox isolation, a refusal of Viewer execution mode, and a deploy-time check of every runnable the app can reach. BuilderKind arrives with the second kind and is a mandatory argument, so no gate can mean "either" by accident: one that did would authorize an app write on a flows-only workspace. The only surfaces reading .any() are the ones that are genuinely not per-kind, the seat a membership costs and listing drafts. validate_operator_composed_app checks two surfaces rather than one, because they are not the same list: the policy's triggerables and the runnableByPath entries in the app value, which is what the deployed bundle actually resolves against. Every colon in a triggerable key is a possible split, since execute_component looks up "<component>:<path>" with an unrestricted component string. create_app_raw_source stays closed: it compiles caller-supplied sources with a bundler job on a worker, which is arbitrary code execution. Builders lose nothing, since the browser and the CLI both bundle locally and deploy through the multipart endpoints. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Dsf6VC4MVLisiEoeQkgbr4 * fix: list a builder's app drafts and refuse inline code in them Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * chore: update ee-repo-ref to 6d15ff957df4dcd5129e1ab217eab1a13af62fbb Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: check a builder app as it will be stored Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: show the home list edit button to builder operators Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat: pause the raw app preview for an app an operator deployed Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: let app builders save a full-code app's deploy settings Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: record the builder app settings path, public apps and the editor preview Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: cache the merged EE seat usage query The EE query now reads audit_partitioned and skips service accounts, so the cached entry for its builder_apps predicate no longer matched and offline EE builds failed with no cached data. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: pause the raw app preview for loaded drafts and unknown authors A draft loaded from an operator ran in the viewer's preview because the pause only checked the app's last deployer, and an author no longer in the workspace read as a developer. The pause now checks a loaded draft's owner, records it in the viewer's draft until the preview is run or deployed, and holds when the author cannot be found. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: name guest apps among the modes a builder may deploy Guest runs as the publisher with no Viewer fallback, like Publisher and Anonymous, and RestrictGuestAppDeployment refuses it per workspace. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: keep a loaded draft's preview paused in AI sessions and chat edits The session draft codec and the chat's draft save dropped loaded_from, so opening a paused operator draft in an AI session, or editing it through the chat, ran it with the viewer's session. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test: pin builder refusals for low-code apps and unreviewed previews Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: keep the operator mark when loading a developer's copy of a draft Loading another user's draft recorded only who saved it, so a developer's copy of an operator draft they never previewed opened unpaused for the next person who loaded it, running the operator's code with their session. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: let builders restore a full-code app from deployment history Restore always sends allow_kind_change, which the builder check refused even for a full-code version restored over a full-code app. Builders now drop the flag, so the locked kind check still refuses an actual conversion. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * style: rustfmt the builder checks in jobs.rs Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: record the builder kind-check race as an accepted risk Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: never trust the owner name a loaded draft carries A loaded draft's loaded_from is written by its owner, so an operator could name a developer, or leave it empty, and the next person to load the draft previewed it unpaused. A loaded draft that carries any mark now stays paused as unattributable; an unmarked one is still checked against its real owner. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: record viewer-supplied resources in builder apps as an accepted risk Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: say that the app right lets operators write frontend code The settings section and the docs told admins builder operators "still cannot write code", but the app right has them write the app's frontend bundle. Only backend code stays out of reach. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: state the raw-source closure once and drop an unreachable preview case Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * perf: look up only the author when deciding whether to pause the preview Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test: pin that the app right alone bills operators as developers Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * chore: point ee-repo-ref at the EE seat test Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor: drop the editor preview pause now that sandboxed apps preview isolated Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: refuse a builder operator's app draft without the sandbox Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: turn the sandbox on when a builder operator opens the app editor Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: refuse builder app drafts naming runnables they cannot read Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: note the editor preview asks before minting a builder app token Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor: trim the builder app check comments to the constraints Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: refuse an app runnable whose runType is not script, flow or hub Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: check builder app draft policy keys like a deploy does Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test: count main's half-seat service accounts in the builder billing test Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor: name the builder policy check for drafts as well as deploys Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: let a builder app draft autosave with a stale triggerables_v2 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: check the script version a builder app would run, not any version Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: refuse viewer mode in a builder app draft as on deploy Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor: hold builder app drafts and deploys to one check Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: say a builder's script may not have finished deploying Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: cap builder app triggerable keys before reading every colon split Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: refuse a builder app draft that runs as someone else Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: conform builder app drafts saved by the AI chat as the editor does Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: record that a deployed builder draft runs as its deployer Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: refuse builder app runnables not written as plain text Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: a builder draft may carry its builder's own run identity Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: offer builders Start with AI only when it opens a session Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: name full-code apps in the builder rights confirmation Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: name only the granted builder rights in the confirmation Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: note a wm_deployers builder may keep another run identity Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: give builders a starter app that calls no missing runnables Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: seed the builder starter under the picker for builders Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: qualify the builder run identity comment for wm_deployers Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: resolve builder app script hashes before opening the RLS transaction Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test: run the builder app tests serially to share the rights cache safely Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: give builders the builder starter on a fresh editor load Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: drop a new-app load overtaken while waiting for the workspace list Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: keep main's 401 for operators previewing without the app right Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * chore: update ee-repo-ref to 5d7726c4979070ee8db4fb9e8014152872a641ba This commit updates the EE repository reference after PR #816 was merged in windmill-ee-private. Previous ee-repo-ref: 7151060bfc02908e97d9344b3820ee33a4951322 New ee-repo-ref: 5d7726c4979070ee8db4fb9e8014152872a641ba Automated by sync-ee-ref workflow. * refactor: let the app row's edit button trust its write flag Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * chore: point ee-repo-ref at the merged EE companion Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: note the session raw-app editor does not conform builder drafts yet Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
Windmill CLI
A simple CLI allowing interactions with windmill from the command line.

You can find more information in Windmill Docs
Installation
Install the wmill CLI tool using npm install -g windmill-cli.
Update to the latest version using wmill upgrade.
Workspaces
To get started run wmill workspace add or use the instructions from the
workspace settings.
Running Flows & Scripts
Run a script or flow using wmill flow/script run u/username/path/to/script and
pass any inputs using --data + Inputs specified as a JSON string or a file
using @ <filename> or stdin using @-.
Curl-style syntax using -d @- for stdin or -d @<filename> is also supported.
Flow Steps and Logs will be streamed during execution automatically.
Pushing Resources, Scripts & More
The CLI can push specifications to a windmill instance. See the examples/ folder for formats.
Switch to a different workspace
wmill workspace switch <workspace_name>
Sync a workspace
Pull
wmill sync pull
Push
wmill sync push
We recommend using the --yaml option to use yaml instead of json as the encoding format. Yaml will be made the default soon.
Pushing individual files
You can push individual resources using
wmill <type> push <file_name> \<remote_name\>. This does not require a special
folder layout or file name, as this is given at runtime.
Listing
All commands support listing by just not providing a subcommand, ie
wmill script will result in a list of scripts. Some allow additional options,
learn about this by specifying --help.
User Management
You can add & remove users via wmill user add/remove, and list them using
wmill user
Pulling
You can pull the entire workspace using wmill pull
Completion
The CLI comes with completions out of the box via wmill completions <shell>.
(Via cliffy)
Bash
To enable bash completions add the following line to your ~/.bashrc:
source <(wmill completions bash)
Fish
To enable fish completions add the following line to your
~/.config/fish/config.fish:
source (wmill completions fish | psub)
Zsh
To enable zsh completions add the following line to your ~/.zshrc:
source <(wmill completions zsh)
Development
AI Guidance Variants
wmill init can now materialize alternate AI guidance bundles without changing
the generated defaults in the repo, but this is exposed as internal env-var
overrides rather than public CLI flags.
Examples:
WMILL_INIT_AI_SKILLS_SOURCE=/path/to/custom/skills wmill init --use-default
WMILL_INIT_AI_SKILLS_SOURCE=/path/to/custom/skills WMILL_INIT_AI_AGENTS_SOURCE=/path/to/AGENTS.md wmill init --use-default
WMILL_INIT_AI_SKILLS_SOURCE=/path/to/custom/skills WMILL_INIT_AI_CLAUDE_SOURCE=/path/to/CLAUDE.md wmill init --use-default
This is the same guidance-writing path used by the benchmark CLI under
ai_evals/, so the benchmark harness and wmill init now generate the same
project guidance shape:
AGENTS.mdCLAUDE.md.agents/skills/*.claude/skills/*
windmill-yaml-validator
wmill lint imports the sibling windmill-yaml-validator package from source rather than
from npm, so its schemas always match the OpenAPI specs of the current checkout. bun install regenerates them through this package's preinstall script; run it again after
editing openflow.openapi.yaml or backend/windmill-api/openapi.yaml:
npm --prefix ../windmill-yaml-validator run gen
Running Tests
Prerequisites:
- PostgreSQL running locally (default:
postgres://postgres:changeme@localhost:5432) - Rust toolchain installed
Run tests locally (full features):
bun test test/
Run tests in CI mode (minimal features, skips EE tests):
CI_MINIMAL_FEATURES=true bun test test/
| Variable | Description |
|---|---|
CI_MINIMAL_FEATURES |
Set to true to skip EE-dependent tests |
DATABASE_URL |
PostgreSQL connection string |
EE_LICENSE_KEY |
Enterprise license key for EE features |
