* Add AWS Bedrock session token support in API and worker
* Add Bedrock auth mode integration tests for AI agents
* Split Bedrock integration test env vars for IAM and session creds
* cleaning
* Add masked Bedrock bearer-token debug info logs
* Revert "Add masked Bedrock bearer-token debug info logs"
This reverts commit 6b2fc5e7c2d5b1c6db81f416a4439941a084108c.
* cleaning
* fix: replace leftover common:: references in dependency_map test
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add missing deno_core/mcp features and gate dead code in permissions test
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
The recent refactor of CloseButton (from on:close component events to
onClick prop) broke tag removal in MultiSelect/TagsToListenTo. The
refactor changed on:pointerdown (component event) to onPointerdown
(native DOM event), which stopped native pointerdown propagation and
broke the drag tracking in DraggableTags, causing the dropdown to open
on every close button click.
Reverts CloseButton and all callers back to using createEventDispatcher
and on:close.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* feat(ai): support 1M context window for Anthropic resources
Add `enable_1m_context` boolean field to Anthropic resource configuration.
When enabled (and not using Vertex AI), sends the `anthropic-beta: context-1m-2025-08-07`
header in both the API proxy layer and the AI agent worker layer.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(ai): add serde alias for enable_1M_context DB field name
The resource_type schema uses `enable_1M_context` (uppercase M) but
serde only matched `enable_1m_context` and `enable1mContext`, causing
the field to always deserialize as false.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
When eval_timeout_quickjs hits the timeout (typically due to slow DB
result retrieval), retry up to 2 more times with a 5s interval between
attempts. Non-timeout errors are returned immediately without retry.
Also extract the eval timeout duration as EVAL_TIMEOUT_MS const (set to
20000ms, up from 10000ms) in windmill-jseval.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Alexander Petric <alpetric@users.noreply.github.com>
* feat: add prompt caching support for Anthropic API
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* exclude vertex
---------
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
* feat: add workspace search and runnable details tools to navigator mode
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: correct uFuzzy search result indexing in workspace search
uFuzzy.search() returns [idxs, info, order] where order contains indices
into idxs, not into the original haystack. The code was using order values
directly as array indices, returning wrong results.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* refactor: mutualize search_workspace and get_runnable_details tools
- Move search_workspace tool def + implementation into shared.ts as
createSearchWorkspaceTool() factory, used by navigator and flow modes
- Move get_runnable_details tool into shared.ts as
createGetRunnableDetailsTool() factory, used by navigator, flow, and
script modes
- Replace flow mode's scripts-only search_scripts with search_workspace
that searches both scripts and flows
- Add search_workspace and get_runnable_details to script mode
- Remove duplicated WorkspaceScriptsSearch class from flow/core.ts
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add get_runnable_details to flow mode system prompt
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add hard limit on runnable content passed to AI context
Truncate script content and flow value at 20k chars in
get_runnable_details to avoid flooding the context window.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: make search_workspace type param required for strict schema
OpenAI strict mode requires all properties in required array. Make type
a required enum ('all', 'scripts', 'flows') instead of optional.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* cleaning
* nit
* cleaning
* refactor: use shared createSearchWorkspaceTool in app mode
Replace app mode's local list_workspace_runnables tool with the shared
createSearchWorkspaceTool() factory, consistent with navigator, flow,
and script modes.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* search by keyword
* cleaning
* fix: document search_workspace and get_runnable_details in script mode system prompt
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add get_runnable_details tool to app mode
Without it, the AI can find scripts/flows but can't inspect their
schema/content when configuring backend runnables with correct inputs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: race condition in WorkspaceRunnablesSearch workspace caching
Track scriptsWorkspace and flowsWorkspace separately instead of a single
shared workspace field. Previously, initScripts could update the shared
workspace field, causing initFlows to skip re-fetching when the workspace
changed (it saw the workspace already matched), returning stale data.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Add new BENCHMARK_KIND variants (sequentialflow, scriptlogs, concurrencylimit,
concurrencykey, mixed, mixed_no_cc) for targeted performance testing. Fix shared
iteration counting across workers using a global atomic counter. Add job_perms
inserts and queue diagnostics for benchmark mode.
Move db connection setup to dedicated module and drop the initial connection pool
before creating the main one, preventing connection starvation when PostgreSQL
max_connections is low.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
