feat: allow disabling run time cred check

[wylswz]

Important

1. Make sure you have read our contribution guidelines
2. Ensure there is an associated issue and you have been assigned to it
3. Use the correct syntax to link this PR: Fixes #<issue number>.

Summary

Credential policy check at run-time introduces degradation on TTFE, so we provide an option to disabling it for latency critical users.

The validation is still performed while publishing the workflow.

Screenshots

Before	After
...	...

Checklist

• This change requires a documentation update, included: Dify Document
• I understand that this PR may be closed in case there was no previous discussion or issues. (This doesn't apply to typos!)
• I've added a test for each change that was introduced, and I tried as much as possible to make a single atomic change.
• I've updated the documentation accordingly.
• I ran make lint && make type-check (backend) and cd web && pnpm exec vp staged (frontend) to appease the lint gods

[github-actions]

Pyrefly Diff

base → PR

--- /tmp/pyrefly_base.txt	2026-05-13 12:34:15.651349885 +0000
+++ /tmp/pyrefly_pr.txt	2026-05-13 12:34:03.547259206 +0000
@@ -28,6 +28,8 @@
    --> core/app/task_pipeline/easy_ui_based_generate_task_pipeline.py:321:29
 ERROR Argument `list[Unknown]` is not assignable to parameter `value` with type `SQLCoreOperations[dict[str, Any]] | dict[str, Any]` in function `sqlalchemy.orm.base.Mapped.__set__` [bad-argument-type]
    --> core/app/task_pipeline/easy_ui_based_generate_task_pipeline.py:388:27
+ERROR Cannot find module `dify.api.configs` [missing-import]
+ --> core/helper/credential_utils.py:5:1
 ERROR No matching overload found for function `core.model_manager.ModelInstance.invoke_llm` called with arguments: (prompt_messages=list[UserPromptMessage], model_parameters=dict[str, int], stream=Literal[False]) [no-matching-overload]
    --> core/llm_generator/llm_generator.py:113:60
 ERROR No matching overload found for function `core.model_manager.ModelInstance.invoke_llm` called with arguments: (prompt_messages=list[UserPromptMessage], model_parameters=dict[str, float | int] | dict[str, object] | dict[Unknown, Unknown], stop=list[str] | list[Unknown], stream=Literal[False]) [no-matching-overload]

[github-actions]

Pyrefly Type Coverage

Metric	Base	PR	Delta
Type coverage	0.00%	43.69%	+43.69%
Strict coverage	0.00%	43.22%	+43.22%
Typed symbols	0	22,031	+22,031
Untyped symbols	0	28,701	+28,701
Modules	0	2552	+2,552

[wylswz]

@dosubot review

[dosubot]

I appreciate you tagging me, but I don't perform traditional code review (assessing PRs for bugs, design, style, performance, or security).

However, I can help in other ways:

• Research the affected code areas or related documentation
• Answer questions about how the credential check logic works today
• Suggest documentation updates if this feature needs user-facing docs

Let me know if any of those would be useful!

To reply, just mention @dosu.

---

Share context across your team and agents. Try Dosu.

[GareArc]

lgtm