Automate browser based workflows with AI
Python aiapiautomationbrowserbrowser-automationcomputergptllmplaywrightpowerautomatepuppeteerpythonrpaseleniumvisionworkflow
fix: teach script reviewer to detect hardcoded per-run values in selectors (#5325)
pedrohsdbeae6622d15ac
diff --git a/skyvern-frontend/src/api/types.ts b/skyvern-frontend/src/api/types.tsindex da99af5c9..b8d1a0626 100644--- a/skyvern-frontend/src/api/types.ts+++ b/skyvern-frontend/src/api/types.ts@@ -426,7 +426,7 @@ export type WorkflowRunApiResponse = {finished_at: string | null; // ISO 8601modified_at: string;proxy_location: ProxyLocation | null;- script_run: { ai_fallback_triggered: boolean } | null;+ script_run: boolean | null;status: Status;title?: string;trigger_type?: TriggerType | null;diff --git a/skyvern-frontend/src/hooks/useRunsQuery.ts b/skyvern-frontend/src/hooks/useRunsQuery.tsindex fc6ebdf30..2ea9cb500 100644--- a/skyvern-frontend/src/hooks/useRunsQuery.ts+++ b/skyvern-frontend/src/hooks/useRunsQuery.ts@@ -3,13 +3,19 @@ import { useCredentialGetter } from "@/hooks/useCredentialGetter";import { useQuery } from "@tanstack/react-query";import { Status, Task, TriggerType, WorkflowRunApiResponse } from "@/api/types";+type QueryReturnType = Array<Task | WorkflowRunApiResponse>;+type UseQueryOptions = Omit<+ Parameters<typeof useQuery<QueryReturnType>>[0],+ "queryKey" | "queryFn"+>;+type Props = {page?: number;pageSize?: number;statusFilters?: Array<Status>;triggerTypeFilters?: Array<TriggerType>;search?: string;-};+} & UseQueryOptions;function useRunsQuery({page = 1,@@ -17,11 +23,9 @@ function useRunsQuery({statusFilters,triggerTypeFilters,search,- ...queryOptions}: Props) {const credentialGetter = useCredentialGetter();return useQuery<Array<Task | WorkflowRunApiResponse>>({- ...queryOptions,queryKey: ["runs",{ statusFilters, triggerTypeFilters },diff --git a/skyvern-frontend/src/routes/history/RunHistory.tsx b/skyvern-frontend/src/routes/history/RunHistory.tsxindex de75b8d25..245adc039 100644--- a/skyvern-frontend/src/routes/history/RunHistory.tsx+++ b/skyvern-frontend/src/routes/history/RunHistory.tsx@@ -279,7 +279,7 @@ function RunHistory() {const workflowTitle = (<div className="flex items-center gap-2"><span className="truncate">{run.workflow_title ?? ""}</span>- {run.script_run != null && (+ {run.script_run === true && (<Tip content="Ran with code"><LightningBoltIcon className="text-[gold]" /></Tip>diff --git a/skyvern-frontend/src/routes/workflows/WorkflowPage.tsx b/skyvern-frontend/src/routes/workflows/WorkflowPage.tsxindex 543dec952..80d36929e 100644--- a/skyvern-frontend/src/routes/workflows/WorkflowPage.tsx+++ b/skyvern-frontend/src/routes/workflows/WorkflowPage.tsx@@ -212,7 +212,7 @@ function WorkflowPage() {) : (workflowRuns?.map((workflowRun) => {const workflowRunId =- workflowRun.script_run != null ? (+ workflowRun.script_run === true ? (<div className="flex items-center gap-2"><Tip content="Ran with code"><LightningBoltIcon className="text-[gold]" />diff --git a/skyvern/forge/prompts/skyvern/script-reviewer-extraction.j2 b/skyvern/forge/prompts/skyvern/script-reviewer-extraction.j2index 07c1b64be..f47428590 100644--- a/skyvern/forge/prompts/skyvern/script-reviewer-extraction.j2+++ b/skyvern/forge/prompts/skyvern/script-reviewer-extraction.j2@@ -34,6 +34,14 @@ These are the KNOWN parameter names for `context.parameters[...]`:{% endfor %}For fields not covered by these parameters, use `ai='proactive'` with a descriptive prompt (see Rule 8b).{% endif %}+{% if run_parameter_values %}++## Current Run Parameter Values+These are the ACTUAL values for this run. If any of these appear (even partially) in a selector or hardcoded string, that value is per-run and MUST be made dynamic — use `context.parameters['key']` or `ai='proactive'` with a prompt referencing the parameter.+{% for key, value in run_parameter_values.items() %}+- `{{ key }}` = `{{ value[:80] }}{% if value|length > 80 %}…{% endif %}`+{% endfor %}+{% endif %}## Existing Cached Code (DO NOT REMOVE ANY EXISTING PATHS)```python@@ -93,9 +101,11 @@ The following failures occurred in PREVIOUS runs. Learn from these — do NOT re- **Error**: {{ h.error_message or "N/A" }}- **Reviewer fix applied**: {{ h.reviewer_output or "No fix recorded" }}- **Did the AI fallback succeed?**: {{ "Yes" if h.fallback_succeeded else ("No" if h.fallback_succeeded is sameas false else "Unknown") }}+{% if h.run_parameters %}- **Run parameters**: {% for k, v in h.run_parameters.items() %}`{{ k }}={{ v[:60] }}{% if v|length > 60 %}…{% endif %}`{% if not loop.last %}, {% endif %}{% endfor %}+{% endif %}{% endfor %}-**Key takeaway:** If the same error appears above multiple times, the previous fixes did NOT solve it. Try a fundamentally different approach.+**Key takeaway:** If the same error appears above multiple times, the previous fixes did NOT solve it. Try a fundamentally different approach. **If different past episodes show different values for the same parameter**, the block handles per-run data — any hardcoded reference to those values must be dynamic.{% endif %}{% if user_instructions %}diff --git a/skyvern/forge/prompts/skyvern/script-reviewer-form-filling.j2 b/skyvern/forge/prompts/skyvern/script-reviewer-form-filling.j2index ae7eaf3b1..45c325d46 100644--- a/skyvern/forge/prompts/skyvern/script-reviewer-form-filling.j2+++ b/skyvern/forge/prompts/skyvern/script-reviewer-form-filling.j2@@ -40,6 +40,14 @@ These are the KNOWN parameter names for `context.parameters[...]`:{% endfor %}For fields not covered by these parameters, use `"ai": "proactive"` with a descriptive prompt (see Rule 9b).{% endif %}+{% if run_parameter_values %}++## Current Run Parameter Values+These are the ACTUAL values for this run. If any of these appear (even partially) in a FIELD_MAP value or hardcoded string, that value is per-run and MUST use `context.parameters['key']` instead.+{% for key, value in run_parameter_values.items() %}+- `{{ key }}` = `{{ value[:80] }}{% if value|length > 80 %}…{% endif %}`+{% endfor %}+{% endif %}## Existing Cached Code```python@@ -96,9 +104,11 @@ The following failures occurred in PREVIOUS runs. Learn from these — do NOT re- **Error**: {{ h.error_message or "N/A" }}- **Reviewer fix applied**: {{ h.reviewer_output or "No fix recorded" }}- **Did the AI fallback succeed?**: {{ "Yes" if h.fallback_succeeded else ("No" if h.fallback_succeeded is sameas false else "Unknown") }}+{% if h.run_parameters %}- **Run parameters**: {% for k, v in h.run_parameters.items() %}`{{ k }}={{ v[:60] }}{% if v|length > 60 %}…{% endif %}`{% if not loop.last %}, {% endif %}{% endfor %}+{% endif %}{% endfor %}-**Key takeaway:** If the same error appears above multiple times, the previous fixes did NOT solve it. Try a fundamentally different approach.+**Key takeaway:** If the same error appears above multiple times, the previous fixes did NOT solve it. Try a fundamentally different approach. **If different past episodes show different values for the same parameter**, the block handles per-run data — any hardcoded reference to those values must use `context.parameters['key']` instead.{% endif %}{% if user_instructions %}diff --git a/skyvern/forge/prompts/skyvern/script-reviewer.j2 b/skyvern/forge/prompts/skyvern/script-reviewer.j2index e5d1cd253..1d4a2dad3 100644--- a/skyvern/forge/prompts/skyvern/script-reviewer.j2+++ b/skyvern/forge/prompts/skyvern/script-reviewer.j2@@ -30,6 +30,21 @@ You are a script reviewer for a browser automation system. Your job is to update11. **FOR-LOOP BLOCKS**: When code runs inside a for-loop (e.g. by iterating over extracted items), every `page.click()`, `page.goto()`, or `page.download_file()` call MUST use per-iteration data to target the specific item. Never use a static selector or generic prompt that would resolve to the same element on every iteration.- **File download blocks**: Use `context.download_selector()` which builds a CSS selector (e.g. `a[href*="filename.pdf"]`) from URL-like fields in the current loop value. Combine with `context.prompt` for AI fallback: `await page.click(selector=context.download_selector(), prompt=context.prompt, ai="fallback")`. If `download_selector()` returns `None` (no URL in loop value), fall back to a text-based selector using `context.loop_value`.- **Other blocks**: Use `context.loop_value` (dict) to build selectors. Prefer `page.goto(context.loop_value["url"])` for direct URLs, or include the item's title/identifier in the selector: `selector=f'a:has-text("{context.loop_value["title"]}")'`.+12. **NEVER HARDCODE PER-RUN VALUES IN SELECTORS OR PROMPTS**. If a selector or prompt contains a value that varies per run (a person's name, an ID, an address, a company name), it MUST be dynamic — never a literal string. This applies even when the page shows an abbreviated form (e.g., "J. Smith") that doesn't exactly match the parameter ("John H. Smith").+ - **How to detect**: Check the **Current Run Parameter Values** and **Past Episode Parameters** sections below. If different episodes show different values for the same parameter (e.g., `full_name: "John Smith"` in one run, `full_name: "Jane Doe"` in another), any selector referencing that name is per-run and must be dynamic.+ - **Wrong** (hardcodes a name that changes every run):+ ```python+ await page.click(selector='a[aria-label="John Smith select for details"]', ai='fallback', prompt='Click provider')+ ```+ - **Correct — option A** (dynamic selector via f-string, with AI fallback for fuzzy matching):+ ```python+ await page.click(selector=f'a:has-text("{context.parameters["full_name"]}")', ai='fallback', prompt=f"Click on the entry for {context.parameters['full_name']}")+ ```+ - **Correct — option B** (AI-only, best when the page abbreviates or reformats the name):+ ```python+ await page.click(ai='proactive', prompt=f"Click on the provider entry that matches {context.parameters['full_name']}")+ ```+ - Use **option A** when the page shows the exact parameter value (fast, no LLM call if selector hits). Use **option B** when the page reformats/abbreviates names — the AI handles fuzzy matching that a static selector cannot.## Block Navigation Goal```@@ -44,6 +59,14 @@ These are the KNOWN parameter names for `context.parameters[...]`:{% endfor %}For fields not covered by these parameters: use Python code if the value is deterministic (see "Deterministic Logic"), or `ai='proactive'` if it requires subjective judgment (see Rule 9c).{% endif %}+{% if run_parameter_values %}++## Current Run Parameter Values+These are the ACTUAL values for this run. If any of these appear (even partially) in a selector or hardcoded string, that value is per-run and MUST be made dynamic (see Rule 12).+{% for key, value in run_parameter_values.items() %}+- `{{ key }}` = `{{ value[:80] }}{% if value|length > 80 %}…{% endif %}`+{% endfor %}+{% endif %}## Existing Cached Code```python@@ -110,9 +133,11 @@ The following failures occurred in PREVIOUS runs and were already reviewed. Lear- **Error**: {{ h.error_message or "N/A" }}- **Reviewer fix applied**: {{ h.reviewer_output or "No fix recorded" }}- **Did the AI fallback succeed?**: {{ "Yes" if h.fallback_succeeded else ("No" if h.fallback_succeeded is sameas false else "Unknown") }}+{% if h.run_parameters %}- **Run parameters**: {% for k, v in h.run_parameters.items() %}`{{ k }}={{ v[:60] }}{% if v|length > 60 %}…{% endif %}`{% if not loop.last %}, {% endif %}{% endfor %}+{% endif %}{% endfor %}-**Key takeaway:** If the same error appears above multiple times, the previous fixes did NOT solve it. Try a fundamentally different approach instead of repeating the same pattern.+**Key takeaway:** If the same error appears above multiple times, the previous fixes did NOT solve it. Try a fundamentally different approach instead of repeating the same pattern. **If different past episodes show different values for the same parameter** (e.g., different names, different IDs), the block handles per-run data — any selector or hardcoded string referencing those values must be dynamic (see Rule 12).{% endif %}{% if user_instructions %}diff --git a/skyvern/forge/sdk/db/agent_db.py b/skyvern/forge/sdk/db/agent_db.pyindex 541d1458c..75a5ef980 100644--- a/skyvern/forge/sdk/db/agent_db.py+++ b/skyvern/forge/sdk/db/agent_db.py@@ -109,8 +109,6 @@ __all__ = ["AgentDB", "ScheduleLimitExceededError"]class AgentDB(BaseAlchemyDB):- _background_tasks: set[asyncio.Task] = set() # noqa: RUF012-def __init__(self, database_string: str, debug_enabled: bool = False, db_engine: AsyncEngine | None = None) -> None:super().__init__(db_engine or _build_engine(database_string))self.debug_enabled = debug_enabled@@ -233,25 +231,7 @@ class AgentDB(BaseAlchemyDB):return await self.tasks.clear_task_failure_reason(*args, **kwargs)async def update_task(self, *args: Any, **kwargs: Any) -> Any:- updated_task = await self.tasks.update_task(*args, **kwargs)-- # Best-effort fire-and-forget write-through to task_runs.- # Mirrors the WorkflowService pattern — cron catches any missed syncs.- status = kwargs.get("status")- if status is not None:- task = asyncio.create_task(- self.workflow_params.sync_task_run_status(- organization_id=updated_task.organization_id or "",- run_id=updated_task.task_id,- status=status.value,- started_at=updated_task.started_at,- finished_at=updated_task.finished_at,- ),- )- self._background_tasks.add(task)- task.add_done_callback(self._background_tasks.discard)-- return updated_task+ return await self.tasks.update_task(*args, **kwargs)async def update_task_2fa_state(self, *args: Any, **kwargs: Any) -> Any:return await self.tasks.update_task_2fa_state(*args, **kwargs)@@ -470,25 +450,25 @@ class AgentDB(BaseAlchemyDB):return await self.workflow_params.retrieve_action_plan(*args, **kwargs)async def create_task_run(self, *args: Any, **kwargs: Any) -> Any:- return await self.workflow_params.create_task_run(*args, **kwargs)+ return await self.tasks.create_task_run(*args, **kwargs)async def update_task_run(self, *args: Any, **kwargs: Any) -> Any:- return await self.workflow_params.update_task_run(*args, **kwargs)+ return await self.tasks.update_task_run(*args, **kwargs)async def sync_task_run_status(self, *args: Any, **kwargs: Any) -> Any:- return await self.workflow_params.sync_task_run_status(*args, **kwargs)+ return await self.tasks.sync_task_run_status(*args, **kwargs)async def update_job_run_compute_cost(self, *args: Any, **kwargs: Any) -> Any:- return await self.workflow_params.update_job_run_compute_cost(*args, **kwargs)+ return await self.tasks.update_job_run_compute_cost(*args, **kwargs)async def cache_task_run(self, *args: Any, **kwargs: Any) -> Any:- return await self.workflow_params.cache_task_run(*args, **kwargs)+ return await self.tasks.cache_task_run(*args, **kwargs)async def get_cached_task_run(self, *args: Any, **kwargs: Any) -> Any:- return await self.workflow_params.get_cached_task_run(*args, **kwargs)+ return await self.tasks.get_cached_task_run(*args, **kwargs)async def get_run(self, *args: Any, **kwargs: Any) -> Any:- return await self.workflow_params.get_run(*args, **kwargs)+ return await self.tasks.get_run(*args, **kwargs)# -- Artifact delegates --diff --git a/skyvern/forge/sdk/db/mixins/tasks.py b/skyvern/forge/sdk/db/mixins/tasks.pyindex 01fef36b7..7f42ee438 100644--- a/skyvern/forge/sdk/db/mixins/tasks.py+++ b/skyvern/forge/sdk/db/mixins/tasks.py@@ -14,13 +14,15 @@ from skyvern.forge.sdk.db.models import (ActionModel,StepModel,TaskModel,+ TaskRunModel,WorkflowRunModel,)from skyvern.forge.sdk.db.utils import convert_to_step, convert_to_task, hydrate_action, serialize_proxy_locationfrom skyvern.forge.sdk.models import Step, StepStatus+from skyvern.forge.sdk.schemas.runs import Runfrom skyvern.forge.sdk.schemas.tasks import OrderBy, SortDirection, Task, TaskStatusfrom skyvern.forge.sdk.utils.sanitization import sanitize_postgres_text-from skyvern.schemas.runs import ProxyLocationInput+from skyvern.schemas.runs import ProxyLocationInput, RunStatus, RunTypefrom skyvern.schemas.steps import AgentStepOutputfrom skyvern.webeye.actions.actions import Action@@ -737,3 +739,197 @@ class TasksMixin:)await session.execute(stmt)await session.commit()++ async def sync_task_run_status(+ self,+ organization_id: str,+ run_id: str,+ status: str,+ started_at: datetime | None = None,+ finished_at: datetime | None = None,+ ) -> None:+ """Best-effort write-through: propagate status from source table to task_runs.++ Does NOT raise if the task_runs row is missing (race at creation time).+ """+ try:+ async with self.Session() as session:+ vals: dict[str, Any] = {"status": status}+ if started_at is not None:+ vals["started_at"] = started_at+ if finished_at is not None:+ vals["finished_at"] = finished_at+ stmt = (+ update(TaskRunModel)+ .where(TaskRunModel.run_id == run_id)+ .where(TaskRunModel.organization_id == organization_id)+ .values(**vals)+ )+ await session.execute(stmt)+ await session.commit()+ except Exception:+ LOG.warning(+ "Best-effort task_run status sync failed",+ run_id=run_id,+ organization_id=organization_id,+ status=status,+ exc_info=True,+ )++ @db_operation("create_task_run")+ async def create_task_run(+ self,+ task_run_type: RunType,+ organization_id: str,+ run_id: str,+ title: str | None = None,+ url: str | None = None,+ url_hash: str | None = None,+ status: RunStatus | None = None,+ workflow_permanent_id: str | None = None,+ parent_workflow_run_id: str | None = None,+ debug_session_id: str | None = None,+ # script_run, started_at, finished_at are intentionally omitted here —+ # they are set via update_task_run() after the run starts/finishes (PRs 2-5).+ ) -> Run:+ searchable_text = " ".join(filter(None, [title, url]))+ async with self.Session() as session:+ task_run = TaskRunModel(+ task_run_type=task_run_type,+ organization_id=organization_id,+ run_id=run_id,+ title=title,+ url=url,+ url_hash=url_hash,+ status=status,+ workflow_permanent_id=workflow_permanent_id,+ parent_workflow_run_id=parent_workflow_run_id,+ debug_session_id=debug_session_id,+ searchable_text=searchable_text or None,+ )+ session.add(task_run)+ await session.commit()+ await session.refresh(task_run)+ return Run.model_validate(task_run)++ @db_operation("update_task_run")+ async def update_task_run(+ self,+ organization_id: str,+ run_id: str,+ title: str | None = None,+ url: str | None = None,+ url_hash: str | None = None,+ status: str | None = None,+ started_at: datetime | None = None,+ finished_at: datetime | None = None,+ ) -> None:+ async with self.Session() as session:+ task_run = (+ await session.scalars(+ select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id)+ )+ ).first()+ if not task_run:+ raise NotFoundError(f"TaskRun {run_id} not found")++ if title is not None:+ task_run.title = title+ if url is not None:+ task_run.url = url+ if url_hash is not None:+ task_run.url_hash = url_hash+ if status is not None:+ task_run.status = status+ if started_at is not None:+ task_run.started_at = started_at+ if finished_at is not None:+ task_run.finished_at = finished_at++ # Recompute searchable_text when title or url changes+ if title is not None or url is not None:+ task_run.searchable_text = " ".join(filter(None, [task_run.title, task_run.url])) or None++ await session.commit()++ @db_operation("update_job_run_compute_cost")+ async def update_job_run_compute_cost(+ self,+ organization_id: str,+ run_id: str,+ instance_type: str | None = None,+ vcpu_millicores: int | None = None,+ memory_mb: int | None = None,+ duration_ms: int | None = None,+ compute_cost: float | None = None,+ ) -> None:+ """Update compute cost metrics for a job run."""+ async with self.Session() as session:+ task_run = (+ await session.scalars(+ select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id)+ )+ ).first()+ if not task_run:+ LOG.warning(+ "TaskRun not found for compute cost update",+ run_id=run_id,+ organization_id=organization_id,+ )+ return++ if instance_type is not None:+ task_run.instance_type = instance_type+ if vcpu_millicores is not None:+ task_run.vcpu_millicores = vcpu_millicores+ if memory_mb is not None:+ task_run.memory_mb = memory_mb+ if duration_ms is not None:+ task_run.duration_ms = duration_ms+ if compute_cost is not None:+ task_run.compute_cost = compute_cost+ await session.commit()++ @db_operation("cache_task_run")+ async def cache_task_run(self, run_id: str, organization_id: str | None = None) -> Run:+ async with self.Session() as session:+ task_run = (+ await session.scalars(+ select(TaskRunModel).filter_by(organization_id=organization_id).filter_by(run_id=run_id)+ )+ ).first()+ if task_run:+ task_run.cached = True+ await session.commit()+ await session.refresh(task_run)+ return Run.model_validate(task_run)+ raise NotFoundError(f"Run {run_id} not found")++ @db_operation("get_cached_task_run")+ async def get_cached_task_run(+ self, task_run_type: RunType, url_hash: str | None = None, organization_id: str | None = None+ ) -> Run | None:+ async with self.Session() as session:+ query = select(TaskRunModel)+ if task_run_type:+ query = query.filter_by(task_run_type=task_run_type)+ if url_hash:+ query = query.filter_by(url_hash=url_hash)+ if organization_id:+ query = query.filter_by(organization_id=organization_id)+ query = query.filter_by(cached=True).order_by(TaskRunModel.created_at.desc())+ task_run = (await session.scalars(query)).first()+ return Run.model_validate(task_run) if task_run else None++ @db_operation("get_run")+ async def get_run(+ self,+ run_id: str,+ organization_id: str | None = None,+ ) -> Run | None:+ async with self.Session() as session:+ query = select(TaskRunModel).filter_by(run_id=run_id)+ if organization_id:+ query = query.filter_by(organization_id=organization_id)+ task_run = (await session.scalars(query)).first()+ return Run.model_validate(task_run) if task_run else Nonediff --git a/skyvern/forge/sdk/db/mixins/workflow_parameters.py b/skyvern/forge/sdk/db/mixins/workflow_parameters.pyindex 5d691d0fa..059233d1e 100644--- a/skyvern/forge/sdk/db/mixins/workflow_parameters.py+++ b/skyvern/forge/sdk/db/mixins/workflow_parameters.py@@ -5,7 +5,7 @@ from datetime import datetime, timedeltafrom typing import TYPE_CHECKING, Anyimport structlog-from sqlalchemy import select, update+from sqlalchemy import selectfrom skyvern.config import settingsfrom skyvern.forge.sdk.db._error_handling import db_operation@@ -24,7 +24,6 @@ from skyvern.forge.sdk.db.models import (OutputParameterModel,TaskGenerationModel,TaskModel,- TaskRunModel,WorkflowCopilotChatMessageModel,WorkflowCopilotChatModel,WorkflowParameterModel,@@ -37,7 +36,6 @@ from skyvern.forge.sdk.db.utils import (hydrate_action,)from skyvern.forge.sdk.schemas.ai_suggestions import AISuggestion-from skyvern.forge.sdk.schemas.runs import Runfrom skyvern.forge.sdk.schemas.task_generations import TaskGenerationfrom skyvern.forge.sdk.schemas.tasks import Task, TaskStatusfrom skyvern.forge.sdk.schemas.workflow_copilot import (@@ -59,7 +57,6 @@ from skyvern.forge.sdk.workflow.models.parameter import (WorkflowParameter,WorkflowParameterType,)-from skyvern.schemas.runs import RunStatus, RunTypefrom skyvern.webeye.actions.actions import Actionif TYPE_CHECKING:@@ -567,197 +564,3 @@ class WorkflowParametersMixin:actions = (await session.scalars(query)).all()return [Action.model_validate(action) for action in actions]-- @db_operation("create_task_run")- async def create_task_run(- self,- task_run_type: RunType,- organization_id: str,- run_id: str,- title: str | None = None,- url: str | None = None,- url_hash: str | None = None,- status: RunStatus | None = None,- workflow_permanent_id: str | None = None,- parent_workflow_run_id: str | None = None,- debug_session_id: str | None = None,- # script_run, started_at, finished_at are intentionally omitted here —- # they are set via update_task_run() after the run starts/finishes (PRs 2-5).- ) -> Run:- searchable_text = " ".join(filter(None, [title, url]))- async with self.Session() as session:- task_run = TaskRunModel(- task_run_type=task_run_type,- organization_id=organization_id,- run_id=run_id,- title=title,- url=url,- url_hash=url_hash,- status=status,- workflow_permanent_id=workflow_permanent_id,- parent_workflow_run_id=parent_workflow_run_id,- debug_session_id=debug_session_id,- searchable_text=searchable_text or None,- )- session.add(task_run)- await session.commit()- await session.refresh(task_run)- return Run.model_validate(task_run)-- @db_operation("update_task_run")- async def update_task_run(- self,- organization_id: str,- run_id: str,- title: str | None = None,- url: str | None = None,- url_hash: str | None = None,- status: str | None = None,- started_at: datetime | None = None,- finished_at: datetime | None = None,- ) -> None:- async with self.Session() as session:- task_run = (- await session.scalars(- select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id)- )- ).first()- if not task_run:- raise NotFoundError(f"TaskRun {run_id} not found")-- if title is not None:- task_run.title = title- if url is not None:- task_run.url = url- if url_hash is not None:- task_run.url_hash = url_hash- if status is not None:- task_run.status = status- if started_at is not None:- task_run.started_at = started_at- if finished_at is not None:- task_run.finished_at = finished_at-- # Recompute searchable_text when title or url changes- if title is not None or url is not None:- task_run.searchable_text = " ".join(filter(None, [task_run.title, task_run.url])) or None-- await session.commit()-- async def sync_task_run_status(- self,- organization_id: str,- run_id: str,- status: str,- started_at: datetime | None = None,- finished_at: datetime | None = None,- ) -> None:- """Best-effort write-through: propagate status from source table to task_runs.-- Does NOT raise if the task_runs row is missing (race at creation time).- """- try:- async with self.Session() as session:- vals: dict[str, Any] = {"status": status}- if started_at is not None:- vals["started_at"] = started_at- if finished_at is not None:- vals["finished_at"] = finished_at- stmt = (- update(TaskRunModel)- .where(TaskRunModel.run_id == run_id)- .where(TaskRunModel.organization_id == organization_id)- .values(**vals)- )- await session.execute(stmt)- await session.commit()- except Exception:- LOG.warning(- "Best-effort task_run status sync failed",- run_id=run_id,- organization_id=organization_id,- status=status,- exc_info=True,- )-- @db_operation("update_job_run_compute_cost")- async def update_job_run_compute_cost(- self,- organization_id: str,- run_id: str,- instance_type: str | None = None,- vcpu_millicores: int | None = None,- memory_mb: int | None = None,- duration_ms: int | None = None,- compute_cost: float | None = None,- ) -> None:- """Update compute cost metrics for a job run."""- async with self.Session() as session:- task_run = (- await session.scalars(- select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id)- )- ).first()- if not task_run:- LOG.warning(- "TaskRun not found for compute cost update",- run_id=run_id,- organization_id=organization_id,- )- return-- if instance_type is not None:- task_run.instance_type = instance_type- if vcpu_millicores is not None:- task_run.vcpu_millicores = vcpu_millicores- if memory_mb is not None:- task_run.memory_mb = memory_mb- if duration_ms is not None:- task_run.duration_ms = duration_ms- if compute_cost is not None:- task_run.compute_cost = compute_cost- await session.commit()-- @db_operation("cache_task_run")- async def cache_task_run(self, run_id: str, organization_id: str | None = None) -> Run:- async with self.Session() as session:- task_run = (- await session.scalars(- select(TaskRunModel).filter_by(organization_id=organization_id).filter_by(run_id=run_id)- )- ).first()- if task_run:- task_run.cached = True- await session.commit()- await session.refresh(task_run)- return Run.model_validate(task_run)- raise NotFoundError(f"Run {run_id} not found")-- @db_operation("get_cached_task_run")- async def get_cached_task_run(- self, task_run_type: RunType, url_hash: str | None = None, organization_id: str | None = None- ) -> Run | None:- async with self.Session() as session:- query = select(TaskRunModel)- if task_run_type:- query = query.filter_by(task_run_type=task_run_type)- if url_hash:- query = query.filter_by(url_hash=url_hash)- if organization_id:- query = query.filter_by(organization_id=organization_id)- query = query.filter_by(cached=True).order_by(TaskRunModel.created_at.desc())- task_run = (await session.scalars(query)).first()- return Run.model_validate(task_run) if task_run else None-- @db_operation("get_run")- async def get_run(- self,- run_id: str,- organization_id: str | None = None,- ) -> Run | None:- async with self.Session() as session:- query = select(TaskRunModel).filter_by(run_id=run_id)- if organization_id:- query = query.filter_by(organization_id=organization_id)- task_run = (await session.scalars(query)).first()- return Run.model_validate(task_run) if task_run else Nonediff --git a/skyvern/forge/sdk/db/mixins/workflow_runs.py b/skyvern/forge/sdk/db/mixins/workflow_runs.pyindex 2c2abe91c..c5b646d7a 100644--- a/skyvern/forge/sdk/db/mixins/workflow_runs.py+++ b/skyvern/forge/sdk/db/mixins/workflow_runs.py@@ -298,7 +298,6 @@ class WorkflowRunsMixin:page: int = 1,page_size: int = 10,status: list[WorkflowRunStatus] | None = None,- trigger_type: list[WorkflowRunTriggerType] | None = None,include_debugger_runs: bool = False,search_key: str | None = None,) -> list[WorkflowRun | Task]:@@ -358,10 +357,6 @@ class WorkflowRunsMixin:if status:workflow_run_query = workflow_run_query.filter(WorkflowRunModel.status.in_(status))- if trigger_type:- workflow_run_query = workflow_run_query.filter(- WorkflowRunModel.trigger_type.in_([t.value for t in trigger_type])- )workflow_run_query = workflow_run_query.order_by(WorkflowRunModel.created_at.desc()).limit(limit)workflow_run_query_result = (await session.execute(workflow_run_query)).all()workflow_runs = [@@ -369,20 +364,16 @@ class WorkflowRunsMixin:for run, title in workflow_run_query_result]- # Tasks don't have trigger_type — skip them when filtering by trigger type- if trigger_type:- tasks: list[Task] = []- else:- task_query = (- select(TaskModel)- .filter(TaskModel.organization_id == organization_id)- .filter(TaskModel.workflow_run_id.is_(None))- )- if status:- task_query = task_query.filter(TaskModel.status.in_(status))- task_query = task_query.order_by(TaskModel.created_at.desc()).limit(limit)- task_query_result = (await session.scalars(task_query)).all()- tasks = [convert_to_task(task, debug_enabled=self.debug_enabled) for task in task_query_result]+ task_query = (+ select(TaskModel)+ .filter(TaskModel.organization_id == organization_id)+ .filter(TaskModel.workflow_run_id.is_(None))+ )+ if status:+ task_query = task_query.filter(TaskModel.status.in_(status))+ task_query = task_query.order_by(TaskModel.created_at.desc()).limit(limit)+ task_query_result = (await session.scalars(task_query)).all()+ tasks = [convert_to_task(task, debug_enabled=self.debug_enabled) for task in task_query_result]runs = workflow_runs + tasksdiff --git a/skyvern/forge/sdk/db/repositories/tasks.py b/skyvern/forge/sdk/db/repositories/tasks.pyindex 889aa0ca8..6e54aa086 100644--- a/skyvern/forge/sdk/db/repositories/tasks.py+++ b/skyvern/forge/sdk/db/repositories/tasks.py@@ -1,5 +1,6 @@from __future__ import annotations+import asynciofrom datetime import datetime, timedelta, timezonefrom typing import Any, Sequence@@ -15,13 +16,15 @@ from skyvern.forge.sdk.db.models import (ActionModel,StepModel,TaskModel,+ TaskRunModel,WorkflowRunModel,)from skyvern.forge.sdk.db.utils import convert_to_step, convert_to_task, hydrate_action, serialize_proxy_locationfrom skyvern.forge.sdk.models import Step, StepStatus+from skyvern.forge.sdk.schemas.runs import Runfrom skyvern.forge.sdk.schemas.tasks import OrderBy, SortDirection, Task, TaskStatusfrom skyvern.forge.sdk.utils.sanitization import sanitize_postgres_text-from skyvern.schemas.runs import ProxyLocationInput+from skyvern.schemas.runs import ProxyLocationInput, RunStatus, RunTypefrom skyvern.schemas.steps import AgentStepOutputfrom skyvern.webeye.actions.actions import Action@@ -29,6 +32,8 @@ LOG = structlog.get_logger()class TasksRepository(BaseRepository):+ _background_tasks: set[asyncio.Task] = set() # noqa: RUF012+@db_operation("create_task")async def create_task(self,@@ -470,6 +475,22 @@ class TasksRepository(BaseRepository):updated_task = await self.get_task(task_id, organization_id=organization_id)if not updated_task:raise NotFoundError("Task not found")++ # Best-effort fire-and-forget write-through to task_runs.+ # Mirrors the WorkflowService pattern — cron catches any missed syncs.+ if status is not None:+ bg = asyncio.create_task(+ self.sync_task_run_status(+ organization_id=updated_task.organization_id or "",+ run_id=updated_task.task_id,+ status=status.value,+ started_at=updated_task.started_at,+ finished_at=updated_task.finished_at,+ ),+ )+ self._background_tasks.add(bg)+ bg.add_done_callback(self._background_tasks.discard)+return updated_taskelse:raise NotFoundError("Task not found")@@ -729,3 +750,197 @@ class TasksRepository(BaseRepository):)await session.execute(stmt)await session.commit()++ async def sync_task_run_status(+ self,+ organization_id: str,+ run_id: str,+ status: str,+ started_at: datetime | None = None,+ finished_at: datetime | None = None,+ ) -> None:+ """Best-effort write-through: propagate status from source table to task_runs.++ Does NOT raise if the task_runs row is missing (race at creation time).+ """+ try:+ async with self.Session() as session:+ vals: dict[str, Any] = {"status": status}+ if started_at is not None:+ vals["started_at"] = started_at+ if finished_at is not None:+ vals["finished_at"] = finished_at+ stmt = (+ update(TaskRunModel)+ .where(TaskRunModel.run_id == run_id)+ .where(TaskRunModel.organization_id == organization_id)+ .values(**vals)+ )+ await session.execute(stmt)+ await session.commit()+ except Exception:+ LOG.warning(+ "Best-effort task_run status sync failed",+ run_id=run_id,+ organization_id=organization_id,+ status=status,+ exc_info=True,+ )++ @db_operation("create_task_run")+ async def create_task_run(+ self,+ task_run_type: RunType,+ organization_id: str,+ run_id: str,+ title: str | None = None,+ url: str | None = None,+ url_hash: str | None = None,+ status: RunStatus | None = None,+ workflow_permanent_id: str | None = None,+ parent_workflow_run_id: str | None = None,+ debug_session_id: str | None = None,+ # script_run, started_at, finished_at are intentionally omitted here —+ # they are set via update_task_run() after the run starts/finishes (PRs 2-5).+ ) -> Run:+ searchable_text = " ".join(filter(None, [title, url]))+ async with self.Session() as session:+ task_run = TaskRunModel(+ task_run_type=task_run_type,+ organization_id=organization_id,+ run_id=run_id,+ title=title,+ url=url,+ url_hash=url_hash,+ status=status,+ workflow_permanent_id=workflow_permanent_id,+ parent_workflow_run_id=parent_workflow_run_id,+ debug_session_id=debug_session_id,+ searchable_text=searchable_text or None,+ )+ session.add(task_run)+ await session.commit()+ await session.refresh(task_run)+ return Run.model_validate(task_run)++ @db_operation("update_task_run")+ async def update_task_run(+ self,+ organization_id: str,+ run_id: str,+ title: str | None = None,+ url: str | None = None,+ url_hash: str | None = None,+ status: str | None = None,+ started_at: datetime | None = None,+ finished_at: datetime | None = None,+ ) -> None:+ async with self.Session() as session:+ task_run = (+ await session.scalars(+ select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id)+ )+ ).first()+ if not task_run:+ raise NotFoundError(f"TaskRun {run_id} not found")++ if title is not None:+ task_run.title = title+ if url is not None:+ task_run.url = url+ if url_hash is not None:+ task_run.url_hash = url_hash+ if status is not None:+ task_run.status = status+ if started_at is not None:+ task_run.started_at = started_at+ if finished_at is not None:+ task_run.finished_at = finished_at++ # Recompute searchable_text when title or url changes+ if title is not None or url is not None:+ task_run.searchable_text = " ".join(filter(None, [task_run.title, task_run.url])) or None++ await session.commit()++ @db_operation("update_job_run_compute_cost")+ async def update_job_run_compute_cost(+ self,+ organization_id: str,+ run_id: str,+ instance_type: str | None = None,+ vcpu_millicores: int | None = None,+ memory_mb: int | None = None,+ duration_ms: int | None = None,+ compute_cost: float | None = None,+ ) -> None:+ """Update compute cost metrics for a job run."""+ async with self.Session() as session:+ task_run = (+ await session.scalars(+ select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id)+ )+ ).first()+ if not task_run:+ LOG.warning(+ "TaskRun not found for compute cost update",+ run_id=run_id,+ organization_id=organization_id,+ )+ return++ if instance_type is not None:+ task_run.instance_type = instance_type+ if vcpu_millicores is not None:+ task_run.vcpu_millicores = vcpu_millicores+ if memory_mb is not None:+ task_run.memory_mb = memory_mb+ if duration_ms is not None:+ task_run.duration_ms = duration_ms+ if compute_cost is not None:+ task_run.compute_cost = compute_cost+ await session.commit()++ @db_operation("cache_task_run")+ async def cache_task_run(self, run_id: str, organization_id: str | None = None) -> Run:+ async with self.Session() as session:+ task_run = (+ await session.scalars(+ select(TaskRunModel).filter_by(organization_id=organization_id).filter_by(run_id=run_id)+ )+ ).first()+ if task_run:+ task_run.cached = True+ await session.commit()+ await session.refresh(task_run)+ return Run.model_validate(task_run)+ raise NotFoundError(f"Run {run_id} not found")++ @db_operation("get_cached_task_run")+ async def get_cached_task_run(+ self, task_run_type: RunType, url_hash: str | None = None, organization_id: str | None = None+ ) -> Run | None:+ async with self.Session() as session:+ query = select(TaskRunModel)+ if task_run_type:+ query = query.filter_by(task_run_type=task_run_type)+ if url_hash:+ query = query.filter_by(url_hash=url_hash)+ if organization_id:+ query = query.filter_by(organization_id=organization_id)+ query = query.filter_by(cached=True).order_by(TaskRunModel.created_at.desc())+ task_run = (await session.scalars(query)).first()+ return Run.model_validate(task_run) if task_run else None++ @db_operation("get_run")+ async def get_run(+ self,+ run_id: str,+ organization_id: str | None = None,+ ) -> Run | None:+ async with self.Session() as session:+ query = select(TaskRunModel).filter_by(run_id=run_id)+ if organization_id:+ query = query.filter_by(organization_id=organization_id)+ task_run = (await session.scalars(query)).first()+ return Run.model_validate(task_run) if task_run else Nonediff --git a/skyvern/forge/sdk/db/repositories/workflow_parameters.py b/skyvern/forge/sdk/db/repositories/workflow_parameters.pyindex 28ad01013..b48a43fc4 100644--- a/skyvern/forge/sdk/db/repositories/workflow_parameters.py+++ b/skyvern/forge/sdk/db/repositories/workflow_parameters.py@@ -5,7 +5,7 @@ from datetime import datetime, timedelta, timezonefrom typing import Anyimport structlog-from sqlalchemy import select, update+from sqlalchemy import selectfrom skyvern.config import settingsfrom skyvern.forge.sdk.db._error_handling import db_operation@@ -26,7 +26,6 @@ from skyvern.forge.sdk.db.models import (OutputParameterModel,TaskGenerationModel,TaskModel,- TaskRunModel,WorkflowCopilotChatMessageModel,WorkflowCopilotChatModel,WorkflowParameterModel,@@ -39,7 +38,6 @@ from skyvern.forge.sdk.db.utils import (hydrate_action,)from skyvern.forge.sdk.schemas.ai_suggestions import AISuggestion-from skyvern.forge.sdk.schemas.runs import Runfrom skyvern.forge.sdk.schemas.task_generations import TaskGenerationfrom skyvern.forge.sdk.schemas.tasks import Task, TaskStatusfrom skyvern.forge.sdk.schemas.workflow_copilot import (@@ -61,7 +59,6 @@ from skyvern.forge.sdk.workflow.models.parameter import (WorkflowParameter,WorkflowParameterType,)-from skyvern.schemas.runs import RunStatus, RunTypefrom skyvern.webeye.actions.actions import ActionLOG = structlog.get_logger()@@ -564,197 +561,3 @@ class WorkflowParametersRepository(BaseRepository):actions = (await session.scalars(query)).all()return [Action.model_validate(action) for action in actions]-- @db_operation("create_task_run")- async def create_task_run(- self,- task_run_type: RunType,- organization_id: str,- run_id: str,- title: str | None = None,- url: str | None = None,- url_hash: str | None = None,- status: RunStatus | None = None,- workflow_permanent_id: str | None = None,- parent_workflow_run_id: str | None = None,- debug_session_id: str | None = None,- # script_run, started_at, finished_at are intentionally omitted here —- # they are set via update_task_run() after the run starts/finishes (PRs 2-5).- ) -> Run:- searchable_text = " ".join(filter(None, [title, url]))- async with self.Session() as session:- task_run = TaskRunModel(- task_run_type=task_run_type,- organization_id=organization_id,- run_id=run_id,- title=title,- url=url,- url_hash=url_hash,- status=status,- workflow_permanent_id=workflow_permanent_id,- parent_workflow_run_id=parent_workflow_run_id,- debug_session_id=debug_session_id,- searchable_text=searchable_text or None,- )- session.add(task_run)- await session.commit()- await session.refresh(task_run)- return Run.model_validate(task_run)-- @db_operation("update_task_run")- async def update_task_run(- self,- organization_id: str,- run_id: str,- title: str | None = None,- url: str | None = None,- url_hash: str | None = None,- status: str | None = None,- started_at: datetime | None = None,- finished_at: datetime | None = None,- ) -> None:- async with self.Session() as session:- task_run = (- await session.scalars(- select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id)- )- ).first()- if not task_run:- raise NotFoundError(f"TaskRun {run_id} not found")-- if title is not None:- task_run.title = title- if url is not None:- task_run.url = url- if url_hash is not None:- task_run.url_hash = url_hash- if status is not None:- task_run.status = status- if started_at is not None:- task_run.started_at = started_at- if finished_at is not None:- task_run.finished_at = finished_at-- # Recompute searchable_text when title or url changes- if title is not None or url is not None:- task_run.searchable_text = " ".join(filter(None, [task_run.title, task_run.url])) or None-- await session.commit()-- async def sync_task_run_status(- self,- organization_id: str,- run_id: str,- status: str,- started_at: datetime | None = None,- finished_at: datetime | None = None,- ) -> None:- """Best-effort write-through: propagate status from source table to task_runs.-- Does NOT raise if the task_runs row is missing (race at creation time).- """- try:- async with self.Session() as session:- vals: dict[str, Any] = {"status": status}- if started_at is not None:- vals["started_at"] = started_at- if finished_at is not None:- vals["finished_at"] = finished_at- stmt = (- update(TaskRunModel)- .where(TaskRunModel.run_id == run_id)- .where(TaskRunModel.organization_id == organization_id)- .values(**vals)- )- await session.execute(stmt)- await session.commit()- except Exception:- LOG.warning(- "Best-effort task_run status sync failed",- run_id=run_id,- organization_id=organization_id,- status=status,- exc_info=True,- )-- @db_operation("update_job_run_compute_cost")- async def update_job_run_compute_cost(- self,- organization_id: str,- run_id: str,- instance_type: str | None = None,- vcpu_millicores: int | None = None,- memory_mb: int | None = None,- duration_ms: int | None = None,- compute_cost: float | None = None,- ) -> None:- """Update compute cost metrics for a job run."""- async with self.Session() as session:- task_run = (- await session.scalars(- select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id)- )- ).first()- if not task_run:- LOG.warning(- "TaskRun not found for compute cost update",- run_id=run_id,- organization_id=organization_id,- )- return-- if instance_type is not None:- task_run.instance_type = instance_type- if vcpu_millicores is not None:- task_run.vcpu_millicores = vcpu_millicores- if memory_mb is not None:- task_run.memory_mb = memory_mb- if duration_ms is not None:- task_run.duration_ms = duration_ms- if compute_cost is not None:- task_run.compute_cost = compute_cost- await session.commit()-- @db_operation("cache_task_run")- async def cache_task_run(self, run_id: str, organization_id: str | None = None) -> Run:- async with self.Session() as session:- task_run = (- await session.scalars(- select(TaskRunModel).filter_by(organization_id=organization_id).filter_by(run_id=run_id)- )- ).first()- if task_run:- task_run.cached = True- await session.commit()- await session.refresh(task_run)- return Run.model_validate(task_run)- raise NotFoundError(f"Run {run_id} not found")-- @db_operation("get_cached_task_run")- async def get_cached_task_run(- self, task_run_type: RunType, url_hash: str | None = None, organization_id: str | None = None- ) -> Run | None:- async with self.Session() as session:- query = select(TaskRunModel)- if task_run_type:- query = query.filter_by(task_run_type=task_run_type)- if url_hash:- query = query.filter_by(url_hash=url_hash)- if organization_id:- query = query.filter_by(organization_id=organization_id)- query = query.filter_by(cached=True).order_by(TaskRunModel.created_at.desc())- task_run = (await session.scalars(query)).first()- return Run.model_validate(task_run) if task_run else None-- @db_operation("get_run")- async def get_run(- self,- run_id: str,- organization_id: str | None = None,- ) -> Run | None:- async with self.Session() as session:- query = select(TaskRunModel).filter_by(run_id=run_id)- if organization_id:- query = query.filter_by(organization_id=organization_id)- task_run = (await session.scalars(query)).first()- return Run.model_validate(task_run) if task_run else Nonediff --git a/skyvern/forge/sdk/db/repositories/workflow_runs.py b/skyvern/forge/sdk/db/repositories/workflow_runs.pyindex 8b5b5103c..a8acb8ac3 100644--- a/skyvern/forge/sdk/db/repositories/workflow_runs.py+++ b/skyvern/forge/sdk/db/repositories/workflow_runs.py@@ -307,7 +307,6 @@ class WorkflowRunsRepository(BaseRepository):page: int = 1,page_size: int = 10,status: list[WorkflowRunStatus] | None = None,- trigger_type: list[WorkflowRunTriggerType] | None = None,include_debugger_runs: bool = False,search_key: str | None = None,) -> list[WorkflowRun | Task]:@@ -367,10 +366,6 @@ class WorkflowRunsRepository(BaseRepository):if status:workflow_run_query = workflow_run_query.filter(WorkflowRunModel.status.in_(status))- if trigger_type:- workflow_run_query = workflow_run_query.filter(- WorkflowRunModel.trigger_type.in_([t.value for t in trigger_type])- )workflow_run_query = workflow_run_query.order_by(WorkflowRunModel.created_at.desc()).limit(limit)workflow_run_query_result = (await session.execute(workflow_run_query)).all()workflow_runs = [@@ -378,20 +373,16 @@ class WorkflowRunsRepository(BaseRepository):for run, title in workflow_run_query_result]- # Tasks don't have trigger_type — skip them when filtering by trigger type- if trigger_type:- tasks: list[Task] = []- else:- task_query = (- select(TaskModel)- .filter(TaskModel.organization_id == organization_id)- .filter(TaskModel.workflow_run_id.is_(None))- )- if status:- task_query = task_query.filter(TaskModel.status.in_(status))- task_query = task_query.order_by(TaskModel.created_at.desc()).limit(limit)- task_query_result = (await session.scalars(task_query)).all()- tasks = [convert_to_task(task, debug_enabled=self.debug_enabled) for task in task_query_result]+ task_query = (+ select(TaskModel)+ .filter(TaskModel.organization_id == organization_id)+ .filter(TaskModel.workflow_run_id.is_(None))+ )+ if status:+ task_query = task_query.filter(TaskModel.status.in_(status))+ task_query = task_query.order_by(TaskModel.created_at.desc()).limit(limit)+ task_query_result = (await session.scalars(task_query)).all()+ tasks = [convert_to_task(task, debug_enabled=self.debug_enabled) for task in task_query_result]runs = workflow_runs + tasksdiff --git a/skyvern/forge/sdk/routes/agent_protocol.py b/skyvern/forge/sdk/routes/agent_protocol.pyindex 74b454ac1..4620fddf0 100644--- a/skyvern/forge/sdk/routes/agent_protocol.py+++ b/skyvern/forge/sdk/routes/agent_protocol.py@@ -39,7 +39,7 @@ from skyvern.forge.sdk.core import skyvern_contextfrom skyvern.forge.sdk.core.curl_converter import curl_to_http_request_block_paramsfrom skyvern.forge.sdk.core.permissions.permission_checker_factory import PermissionCheckerFactoryfrom skyvern.forge.sdk.core.security import generate_skyvern_signature-from skyvern.forge.sdk.db.enums import OrganizationAuthTokenType, WorkflowRunTriggerType+from skyvern.forge.sdk.db.enums import OrganizationAuthTokenTypefrom skyvern.forge.sdk.executor.factory import AsyncExecutorFactoryfrom skyvern.forge.sdk.models import Stepfrom skyvern.forge.sdk.routes.code_samples import (@@ -2120,7 +2120,6 @@ async def get_runs(page: int = Query(1, ge=1),page_size: int = Query(10, ge=1),status: Annotated[list[WorkflowRunStatus] | None, Query()] = None,- trigger_type: Annotated[list[WorkflowRunTriggerType] | None, Query()] = None,search_key: str | None = Query(None,description=(@@ -2139,12 +2138,7 @@ async def get_runs(return []runs = await app.DATABASE.get_all_runs(- current_org.organization_id,- page=page,- page_size=page_size,- status=status,- trigger_type=trigger_type,- search_key=search_key,+ current_org.organization_id, page=page, page_size=page_size, status=status, search_key=search_key)return ORJSONResponse([run.model_dump() for run in runs])diff --git a/skyvern/forge/sdk/routes/scripts.py b/skyvern/forge/sdk/routes/scripts.pyindex 67e554a16..a746150ad 100644--- a/skyvern/forge/sdk/routes/scripts.py+++ b/skyvern/forge/sdk/routes/scripts.py@@ -1330,8 +1330,12 @@ async def review_script_with_instructions(workflow_run_id=data.workflow_run_id,)for wf_param, run_param in run_param_tuples:- if isinstance(run_param.value, str) and run_param.value:- run_parameter_values[wf_param.key] = run_param.value+ if (+ run_param.value is not None+ and str(run_param.value).strip()+ and not wf_param.parameter_type.is_secret_or_credential()+ ):+ run_parameter_values[wf_param.key] = str(run_param.value)except Exception:LOG.warning("Failed to load run parameter values", exc_info=True)diff --git a/skyvern/forge/sdk/workflow/service.py b/skyvern/forge/sdk/workflow/service.pyindex 4cf54cdf2..d20a27c79 100644--- a/skyvern/forge/sdk/workflow/service.py+++ b/skyvern/forge/sdk/workflow/service.py@@ -3655,12 +3655,11 @@ class WorkflowService:)# Best-effort fire-and-forget write-through to task_runs table.# Runs off the hot path so workflow status transitions stay fast.- # Hold a reference in _background_tasks so Python doesn't GC the task.- task = asyncio.create_task(+ bg = asyncio.create_task(self._sync_task_run_from_workflow_run(workflow_run, workflow_run_id, status),)- self._background_tasks.add(task)- task.add_done_callback(self._background_tasks.discard)+ self._background_tasks.add(bg)+ bg.add_done_callback(self._background_tasks.discard)return workflow_run@@ -5454,8 +5453,12 @@ class WorkflowService:workflow_run_id=workflow_run.workflow_run_id,)for wf_param, run_param in run_param_tuples:- if isinstance(run_param.value, str) and run_param.value:- run_parameter_values[wf_param.key] = run_param.value+ if (+ run_param.value is not None+ and str(run_param.value).strip()+ and not wf_param.parameter_type.is_secret_or_credential()+ ):+ run_parameter_values[wf_param.key] = str(run_param.value)except Exception:LOG.debug("Failed to load run parameter values for hardcoded-value check", exc_info=True)diff --git a/skyvern/services/script_reviewer.py b/skyvern/services/script_reviewer.pyindex f67c0d0b9..597546f9c 100644--- a/skyvern/services/script_reviewer.py+++ b/skyvern/services/script_reviewer.py@@ -1,5 +1,6 @@from __future__ import annotations+import asyncioimport jsonimport refrom typing import Literal, Sequence@@ -113,6 +114,33 @@ class ScriptReviewer:history_by_block[ep.block_label] = []history_by_block[ep.block_label].append(ep)+ # Batch-load parameter values for historical episodes so the reviewer+ # can detect per-run values (e.g., different provider names across runs).+ # Passed explicitly to _review_block to avoid implicit instance state.+ historical_run_params: dict[str, dict[str, str]] = {}+ if historical_episodes:+ unique_run_ids = list({ep.workflow_run_id for ep in historical_episodes if ep.workflow_run_id})[:20]++ async def _load_run_params(run_id: str) -> tuple[str, dict[str, str]]:+ try:+ param_tuples = await app.DATABASE.get_workflow_run_parameters(workflow_run_id=run_id)+ params = {+ wf_param.key: str(run_param.value)+ for wf_param, run_param in param_tuples+ if run_param.value is not None+ and str(run_param.value).strip()+ and not wf_param.parameter_type.is_secret_or_credential()+ }+ return run_id, params+ except Exception:+ LOG.warning(+ "Failed to load params for historical episode run", workflow_run_id=run_id, exc_info=True+ )+ return run_id, {}++ results = await asyncio.gather(*[_load_run_params(rid) for rid in unique_run_ids])+ historical_run_params = {rid: params for rid, params in results if params}+# Triage failed episodes — skip non-code-fixable failures.# When user provides explicit instructions, skip triage entirely.if user_instructions:@@ -161,6 +189,7 @@ class ScriptReviewer:historical_episodes=history_by_block.get(block_label),run_parameter_values=run_parameter_values,user_instructions=user_instructions,+ historical_run_params=historical_run_params,)if updated_code:updated_blocks[block_label] = updated_code@@ -529,6 +558,7 @@ class ScriptReviewer:run_parameter_values: dict[str, str] | None = None,user_instructions: str | None = None,preloaded_code: str | None = None,+ historical_run_params: dict[str, dict[str, str]] | None = None,) -> str | None:"""Review a single block's fallback episodes and generate updated code."""LOG.info(@@ -604,16 +634,20 @@ class ScriptReviewer:goal_param_keys = set(re.findall(r"\{\{\s*(\w+)\s*\}\}", navigation_goal))parameter_keys = sorted(goal_param_keys | set(all_parameter_keys or []))- # Build historical episode summaries for cross-run context+ # Build historical episode summaries for cross-run context.+ # Include per-run parameter values so the reviewer can detect that+ # different runs had different names/IDs (→ selectors must be dynamic).history_summaries = []for ep in historical_episodes or []:- history_summaries.append(- {- "error_message": ep.error_message,- "reviewer_output": (ep.reviewer_output or "")[:500],- "fallback_succeeded": ep.fallback_succeeded,- }- )+ summary: dict[str, object] = {+ "error_message": ep.error_message,+ "reviewer_output": (ep.reviewer_output or "")[:500],+ "fallback_succeeded": ep.fallback_succeeded,+ }+ ep_params = (historical_run_params or {}).get(ep.workflow_run_id)+ if ep_params:+ summary["run_parameters"] = ep_params+ history_summaries.append(summary)# Build the reviewer promptreviewer_prompt = prompt_engine.load_prompt(@@ -636,6 +670,7 @@ class ScriptReviewer:stale_branches=stale_branch_info,parameter_keys=parameter_keys,historical_episodes=history_summaries,+ run_parameter_values=run_parameter_values,user_instructions=user_instructions,)diff --git a/tests/unit/test_task_run_status_sync.py b/tests/unit/test_task_run_status_sync.pyindex 6b91dec7f..5dde03e33 100644--- a/tests/unit/test_task_run_status_sync.py+++ b/tests/unit/test_task_run_status_sync.py@@ -19,14 +19,14 @@ def mock_session():def agent_db(mock_session):db = AgentDB.__new__(AgentDB)db.Session = MagicMock(return_value=mock_session)- # Set up workflow_params with the same session factory for delegation- from skyvern.forge.sdk.db.repositories.workflow_parameters import WorkflowParametersRepository-- wp = WorkflowParametersRepository.__new__(WorkflowParametersRepository)- wp.Session = MagicMock(return_value=mock_session)- wp.debug_enabled = False- wp._is_retryable_error_fn = None- db.workflow_params = wp+ # Set up tasks repository (sync_task_run_status delegates to self.tasks)+ from skyvern.forge.sdk.db.repositories.tasks import TasksRepository++ tasks = TasksRepository.__new__(TasksRepository)+ tasks.Session = MagicMock(return_value=mock_session)+ tasks.debug_enabled = False+ tasks._is_retryable_error_fn = None+ db.tasks = tasksreturn dbdiff --git a/tests/unit/test_task_runs_sync_activity.py b/tests/unit/test_task_runs_sync_activity.pyindex d7de8d04c..701b84f4d 100644--- a/tests/unit/test_task_runs_sync_activity.py+++ b/tests/unit/test_task_runs_sync_activity.py@@ -117,7 +117,7 @@ def _load_task_runs_sync_activity_module(monkeypatch: pytest.MonkeyPatch):monkeypatch.setitem(sys.modules, "temporalio", temporalio_module)monkeypatch.setitem(sys.modules, "structlog", structlog_module)- module_path = _SOURCE_FILE+ module_path = _repo_root() / "workers" / "cron_worker" / "task_runs_sync_activity.py"spec = importlib.util.spec_from_file_location("test_task_runs_sync_activity_module", module_path)module = importlib.util.module_from_spec(spec)assert spec is not None and spec.loader is not None