* fix: enforce jobs:run scope on job preview and inline endpoints
Preview/inline endpoints (run/preview, run/preview_bundle, run/preview_flow,
run/dynamic_select inline) execute arbitrary request-supplied code but only
checked folder/namespace read access, which is a no-op when path is null. A
token scoped to a specific script/flow could escape its scope and run any
code. Add a jobs:run scope check, matching other arbitrary-execution
endpoints. Advisory GHSA-vxc5-w28p-m9xw.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: scope-check dynamic_select flow branch and inline preview
Address CI review: the dynamic_select Deployed{Flow} branch ran a deployed
flow's dynamic-select code without any scope check (only the Script branch
delegated to a scope-checked handler), and run_inline_preview_script executed
request-supplied code with no in-handler scope check. Add jobs:run:flows:{path}
to the flow branch and jobs:run to inline preview; correct the misleading
comment. Expand regression tests (preview_flow case, assert success for the
broad-token case). Advisory GHSA-vxc5-w28p-m9xw.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test: remove preview scope enforcement test after local validation
The regression test passed locally (3/3) and validated the fix end-to-end;
removed from the PR per maintainer preference. Advisory GHSA-vxc5-w28p-m9xw.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The dev profile's split-debuginfo = "unpacked" is coerced to "packed" on
windows-msvc, so each test-binary link spawns the shared mspdbsrv.exe PDB
type server. With 12 parallel link jobs this races the type-server cap
(LNK1318 "LIMIT (12)") and exhausts the runner disk (LNK1180), recurringly
failing the Windows release CI. CI needs no debug info, so disable PDB
generation for the dev/test profiles in this job only.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: constrain unauthenticated get_public_resource to app_theme resources
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test: remove get_public_resource regression test
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: bind /etc resolver files read-only in nsjail sandboxes
* docs(nsjail): explain why per-file /etc resolver binds are load-bearing
The explicit /etc/hosts, /etc/resolv.conf and /etc/hostname binds look
like removable duplication of the read-only /etc bind above them. They
are not: on Kubernetes those files are separate kubelet bind-mounts on
top of /etc and nsjail's read-only remount is non-recursive, so without
these shadow binds they stay writable and a job can persist cross-tenant
DNS poisoning for the pod lifetime. Comment guards against a future
"dedup cleanup" silently reintroducing the vulnerability.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* docs(nsjail): shorten the load-bearing-bind comment to 3 lines
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
* fix: sidebar menu rendering expanded-but-empty near 768px width
The desktop sidebar branch is gated by JS (`innerWidth < 768`), but its
width was set only via Tailwind `md:` classes (`@media (min-width:768px)`).
`window.innerWidth` rounds fractional viewport widths, so at e.g. 767.8px
JS rounds to 768 and renders the desktop sidebar, while the CSS media
query does not match and no width class applies — leaving the sidebar
shell expanded with no width/content. Drop the now-redundant `md:`
prefix so width tracks the JS branch decision.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: align content offset breakpoint with sidebar JS gate
The sidebar width now follows the JS innerWidth gate, but the main
content left-offset in AiChatLayout still used the `md:` CSS media
query, leaving the two breakpoints out of sync in the same sub-pixel
band. Pass an `isMobile` flag from the layout (mirroring the sidebar's
`innerWidth < 768` condition) and gate the content padding on it with
unprefixed classes so sidebar width and content offset always flip
together.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: reset parent_hash in auto_parent when all versions at path are archived
* test: regression test for auto_parent with all versions archived
---------
Co-authored-by: windmill-internal-app[bot] <1429786+windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
* fix: don't show ALLOW_PRIVATE_AI_BASE_URLS hint for malformed AI base URLs
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* fix: impl std::error::Error for SsrfValidationError for anyhow callers
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
* fix: atomic bundle cache writes to prevent parallel cold-load race
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: trust-but-replace in atomic_publish_dir to never trust a stale partial cache dir
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: simplify atomic_publish_dir and add content-addressed rename-failure fallback
Revert the destroy-then-recreate dir swap (introduced concurrent-publisher
edge cases: spurious Err under a real herd, EACCES masking a stale partial)
back to a single atomic rename. Add the content-addressed exists-fallback to
atomic_write_file_bytes/atomic_copy_file so the loser of a publish race (and
Windows, where rename cannot replace an open/existing destination) treats the
already-published identical cache as success instead of failing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
windmill-parser-ts-asset only recognized writeS3File/loadS3File when
the first arg was a bare 's3://...' string literal. The actual SDK
signature takes an S3Object ({ s3, storage? }) or 's3://bucket/key'
string, which every real script uses, so object-form writes/reads were
never detected as assets. Resolve the S3Object arg the same way the
runtime parseS3Object does, mapping { s3, storage } to s3://<storage>/
<key> and feeding it through parse_asset_syntax so the path matches the
// on s3:///… trigger form. Adds regression tests.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: aggregate wait time should target the true root job, not flow_innermost_root_job
* refactor: reuse get_root_job_id helper for wait-time aggregation
Instead of duplicating the root_job → flow_innermost_root_job →
parent_job fallback chain inline, call the existing get_root_job_id()
helper (the same one used by push_next_flow_job) and filter out the
self-id case so standalone scripts still skip aggregate insertion.
Behaviorally identical, single source of truth.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: windmill-internal-app[bot] <1429786+windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* [ee] fix(nativets): pass tracing-enabled OtelConfig to deno_telemetry::init
* chore: update ee-repo-ref to 19a76a09ffb43649ee19e62d07e8b8a42d78757b
This commit updates the EE repository reference after PR #573 was merged in windmill-ee-private.
Previous ee-repo-ref: daf7ce31f192d64ca6b5602af626ab6453d2aff4
New ee-repo-ref: 19a76a09ffb43649ee19e62d07e8b8a42d78757b
Automated by sync-ee-ref workflow.
---------
Co-authored-by: windmill-internal-app[bot] <1429786+windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
* feat: include service accounts in instance settings users list
Service accounts (workspace-scoped, no password row) now appear in the
superadmin users list with a Bot icon, workspace badge, and a link to
manage them in the workspace settings. Role is locked to Operator.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* chore: update sqlx offline cache
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: use composite key for users each block
Service accounts can share emails across workspaces, so key by email + workspace_id to avoid Svelte each_key_duplicate.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(python): preserve strings containing Infinity/NaN in result JSON
* test(python): add sanity checks for Infinity/NaN in results
* refactor(python): use string-aware regex callback for single-pass cleanup
* refactor(python): compact regex + handle backslash-escape parity
* perf(python): short-circuit cleanup when no Infinity/NaN/NUL in result
* feat: read-only flag on API tokens, orthogonal to scopes
Add a per-token `read_only` boolean set at creation time. When true, the
token can only call HTTP methods classified as Read (GET/HEAD/OPTIONS).
Mutating methods and job-run actions are rejected with 403, regardless of
which scopes are attached. Surfaced as a prominent toggle in the standard
token-creation flow and a discreet `2xs` toggle in MCP mode (where users
often want write access, so we don't bias them toward enabling it).
MCP enforcement: read-only tokens hide all script/flow/hub tools from
`list_tools` and only see endpoint tools whose method is GET, and the
runner rejects `call_tool` on anything mutating.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: review fixes for read-only token flag
- Exempt /api/mcp/* and /mcp/* paths from the read-only middleware check.
MCP transport runs over POST (streamable HTTP / SSE), so otherwise the
middleware would 403 every MCP request before the runner could enforce
read-only at the tool-call level.
- Tighten is_endpoint_read_only to GET only, matching the read_only_hint
that create_endpoint_annotations actually emits.
- Add unit test for check_read_only_for_route covering GET/HEAD/OPTIONS,
mutating methods, and run paths.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* chore: bump ee-repo-ref to read-only-trigger-toggle
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(frontend): make read-only toggle discreet in both modes
Match the MCP-mode treatment in standard mode: text-tertiary, 2xs, shared
"Read-only" label. The tooltip switches per mode so the explanation still
fits the context.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(frontend): gate read-only toggle behind Limit token permissions
The read-only toggle now only shows when the user has limited the token's
scopes (standard mode) or in MCP mode (which always picks an MCP scope).
Turning the limit off also resets read-only so it doesn't silently stick.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(frontend): hide incompatible MCP tools when read-only is on
When the read-only toggle is on in MCP mode:
- Endpoint badges and the custom-mode endpoint MultiSelect filter to GET.
- Already-selected non-GET endpoints are pruned from the scope.
- The scripts/flows preview is replaced with a note explaining they're
hidden (the runner already rejects script/flow runs for read-only).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(frontend): place read-only toggle at top of limited scope area
The previous gate required at least one scope to be picked before the
read-only toggle appeared, which made it look missing while the user was
still building their scope list. Move the toggle inside ScopesPicker:
- Standard mode: sits directly under the "Limit token permissions" toggle
whenever Limit is on, before the scope selector.
- MCP mode: sits at the top of the MCP scope block.
readOnly is now $bindable on ScopesPicker so CreateToken still owns the
value. The auto-reset on un-limit moves into ScopesPicker too.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(frontend): nest read-only toggle inside the scope list card
Place the read-only toggle at the top of the scope list (between the
Selected Scopes summary and the bordered domain list) via a new optional
topSlot snippet on ScopeSelector. Keeps ScopeSelector decoupled from
read-only specifics; ScopesPicker fills the slot.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* chore: update ee-repo-ref to 9bc8160be50b3e57a60daf4e1b71c389a6e02b8a
This commit updates the EE repository reference after PR #571 was merged in windmill-ee-private.
Previous ee-repo-ref: f53d26e6685dfd60bfa67686fbd7358169cfd130
New ee-repo-ref: 9bc8160be50b3e57a60daf4e1b71c389a6e02b8a
Automated by sync-ee-ref workflow.
* fix: address CI review for read-only token flag
- P1 (Codex): narrow the MCP middleware exemption from "any /api/mcp/*"
to just the streamable HTTP transport endpoints
(/api/mcp/gateway, /api/mcp/w/{ws}/{mcp,sse,list_tools}). Without this,
a read-only token could POST /api/mcp/gateway/oauth/server/approve and
mint a follow-on non-read-only MCP token via the OAuth code/token
exchange.
- P2 (Claude/cubic): fix test comment/assertion mismatch — the run-path
assertion now exercises GET (which is what the RUN_PATH_ACTIONS
elevation comment describes) in addition to POST. Add a regression
assertion for /api/mcp/gateway/oauth/server/approve.
- P2 (cubic): short-circuit script/flow/hub-script/resource fetches in
MCP list_tools when read_only is on — they would only be discarded
below, so skipping the DB and resource fan-out is pure win.
- P2 (cubic): when scopes are pre-supplied via the CreateToken prop, the
ScopesPicker isn't rendered, which previously hid the read-only
toggle entirely. Render it next to the pre-supplied scopes display.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
- exists_script_by_path now filters archived = false, matching the
conflict check in create_script_internal. Previously the frontend
blocked creating a new script at a path occupied only by archived
scripts, even though renaming to that same path was allowed.
- Hide the Delete entry in the script details "..." menu unless the
user is admin. The backend delete_script_by_hash already requires
admin, so non-admins would always see an error after clicking.
* perf(dynselect): only retrigger when helper-script args actually change
Parse the inline helper's signature with the existing WASM parser and
restrict the form-arg diff to keys the helper actually consumes. Typing
into unrelated fields no longer queues a dynselect job every second.
Falls back to the previous full-args comparison when the helper is
deployed or parsing fails.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(dynselect): avoid double helper-script fetch on mount
usePromise defaults to loadInit=true, so refresh() ran before the
JobLoader child was bound (firing a no-op pending promise) and the
$effect then fired a second refresh once the bind:this resolved.
Disable loadInit so the effect owns the single first call.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* refactor(dynselect): use parser directly instead of inferArgs
inferArgs mutates a Schema object we never use and goes through a
shared cache; when fed an empty schema for non-main entrypoints the
caller cannot reliably read back the resulting properties. Add
parseEntrypointArgs that just runs the parser and returns the
parameter name Set (or undefined when unknown / unsupported / has
rest args / function not found). DynamicInput uses that and keeps
the previous params in flight while the next parse is computing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat(dynselect): support deployed helpers in smart retrigger
Add getHelperEntrypointArgs which dispatches on HelperScript.source:
inline parses immediately; deployed fetches the script (or the flow's
inline dyn-select code) once and caches per (workspace, kind, path,
entrypoint). Without this the /scripts/get/* run view fell back to
the full-args comparison and still retriggered on unrelated fields.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(dynselect): zero-arg helpers report empty deps, not unknown
Codex review flagged that a valid zero-parameter entrypoint was being
treated as "couldn't determine signature" and falling back to the
full-args comparison. Distinguish "function found with no params" from
"function not found" via the parser's auto_kind field — only the
latter sets it, so empty args + auto_kind=null means a real zero-arg
helper and we return an empty Set (no retrigger on unrelated fields).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(bun): pass --preserve-symlinks on unbundled execution
Bun 1.2/1.3 moved its global package cache to a content-addressed
layout and the installer now creates a single directory symlink from
node_modules/<pkg> to the cache entry. Without --preserve-symlinks,
Bun resolves modules from each file's realpath, so any require/import
inside an installed package walks up from cache_nomount/bun/... and
never finds the sibling deps living under <job_dir>/node_modules/.
This manifested as e.g.
ENOENT while resolving package 'zod/v3' from
'/tmp/windmill/cache_nomount/bun/@langchain/core@1.1.44@@@1/dist/...'
on //nobundling scripts that pull @langchain/core, even though zod is
correctly installed alongside it in node_modules.
The bundled execution path already had --preserve-symlinks since #4132
(needed because we symlink the cached bundle file into the job dir).
The unbundled path didn't, because at the time Bun installed via per-
file hardlinks and the realpath of node_modules entries was the job
dir itself. The Bun installer's layout change made the flag necessary
on the unbundled path as well.
Add the flag to all three unbundled `bun run` invocations:
- nsjail unbundled path
- non-nsjail unbundled path
- dedicated worker (always unbundled)
This also fixes a latent bug on the first run of any bun script that
imports a package whose internals reference siblings (the build_cache
path runs unbundled this round while it builds the bundle for next
time).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test(bun): regression test for nobundling + transitive require resolution
Adds an integration test that mirrors the original failure: a //nobundling
script importing @langchain/core, which (in its CJS internals) does
require('zod/v3'). Before --preserve-symlinks was added to the unbundled
bun run invocations, this failed with:
ENOENT while resolving package 'zod/v3' from
'.../cache_nomount/bun/@langchain/core@<ver>@@@1/dist/runnables/base.js'
The test covers the non-nsjail unbundled path. Reproducibility of the
pre-fix failure depends on Bun's installer choosing the directory-symlink
layout for the node_modules entry (the default on Bun 1.2/1.3+ with the
new content-addressed global cache that produced the user's error).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test(git-sync): regression tests for secondary promotion repos
Adds two integration tests that reproduce the bug where a second
promotion-mode repo's deployment callback was silently dropped via
debounce-key collision, plus the EE ref bump that includes the fix.
Updates the two existing promotion-mode debounce-key tests to expect
the new repo-namespaced key shape.
* test(git-sync): drop redundant distinct-debounce-keys test
The behavior test (`test_two_promotion_repos_both_enqueue_callback`)
already covers the same regression one layer up: if the debounce keys
collide, one callback gets marked skipped, which the behavior test
catches.
* chore: update ee-repo-ref to 7a32388adaa37eb1dd1820b40e140ff1877110f2
This commit updates the EE repository reference after PR #572 was merged in windmill-ee-private.
Previous ee-repo-ref: dbf26f5e4c01c0de536f606679be46eb316aaf31
New ee-repo-ref: 7a32388adaa37eb1dd1820b40e140ff1877110f2
Automated by sync-ee-ref workflow.
---------
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
* feat: publish CLI skills + AGENTS.md to windmill-cli-docs for context7
Auto-generates a public docs snapshot (AGENTS.md, full CLI reference,
all rendered skills) and pushes it to windmill-labs/windmill-cli-docs on
every release tag, so context7 can index Windmill CLI docs.
- generate.py: new --context7-dir flag rendering fully-resolved skills
+ AGENTS.md (extracted from cli/src/guidance/core.ts to avoid drift)
+ cli-commands.md + README.md + manifest.json into a docs-repo checkout.
Preserves .git, .github, LICENSE, context7.json across regenerations.
- publish-cli-docs.yml: GitHub Action on v* tag and workflow_dispatch
that regenerates the docs repo and pushes via the CLI_DOCS_DEPLOY_KEY
SSH deploy key.
* fix: skip tag mirror on workflow_dispatch from non-tag ref
* docs: turn windmill-cli-docs README into a CLI quickstart
* fix: address PR review (target safety, regex anchor, concurrency, tag mirror)
- Refuse to wipe --context7-dir unless empty, has a context7 marker, or
points at the windmill-cli-docs remote (P1, prevents typo blast).
- Anchor AGENTS.md template regex on `generateAgentsMdContent` so adding
other template-returning functions to core.ts can't silently retarget it.
- Decode TS escapes in one pass to avoid order-sensitive mangling.
- Include Windmill version (from version.txt) in manifest.json so each
snapshot is self-describing.
- Add concurrency group on the publish workflow.
- Always mirror version tag on tag pushes, even when content is unchanged,
so the docs repo has a tag for every Windmill release.
- Expand preserve list with .gitignore, .gitattributes, CODEOWNERS.
* fix: validate manifest.json content, not just presence, before wipe
Diego Imbert
Ruben Fiszel
Diego Imbert
windmill-internal-app[bot]
Blossom
centdix
hugocasa