fix(crewai): apply convert_llm and convert_tools in SequentialTaskAgentComponent

[gingeekrishna]

Summary

SequentialTaskAgentComponent.build_agent_and_task passed the raw LangChain LLM handle and tool list directly to crewai.Agent, while the sibling CrewAIAgentComponent.build_output already calls convert_llm() and convert_tools() at construction time.

This asymmetry means any consumer of the SequentialTask.agent output port that is not SequentialCrewComponent receives an Agent whose .llm is a LangChain type rather than a crewai.LLM, causing runtime failures.

Changes:

• Import convert_llm and convert_tools from lfx.base.agents.crewai.crew
• Apply both at Agent(...) construction, matching CrewAIAgentComponent

Fixes #13848

Test plan

• Build a flow using Sequential Task Agent -> Sequential Crew with a non-OpenAI LLM; verify the crew runs without type errors on .llm
• Confirm no regression on existing CrewAI Agent -> Sequential Crew flows

Summary by CodeRabbit

• Bug Fixes
  • Improved agent setup to better handle model and tool inputs, making task execution more reliable.
  • Enhanced compatibility when connecting language models and tools to sequential agent workflows.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[coderabbitai]

Important

Review skipped

Auto incremental reviews are disabled on this repository.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: f0e13643-e160-468a-8d55-bad7dab7f059

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

Walkthrough

SequentialTaskAgentComponent now converts its configured LLM and tools before creating the CrewAI Agent. The task/agent chaining and task field handling remain unchanged.

Changes

Sequential task agent conversion

Layer / File(s)	Summary
LLM and tool conversion src/lfx/src/lfx/components/crewai/sequential_task_agent.py	Adds convert_llm and convert_tools imports, then uses them when building the Agent in build_agent_and_task.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

---

Important

Pre-merge checks failed

Please resolve all errors before merging. Addressing warnings is optional.

❌ Failed checks (1 error, 2 warnings)

Check name	Status	Explanation	Resolution
Test Coverage For New Implementations	❌ Error	Only the component source changed; no new or updated test file covers the convert_llm/convert_tools fix. The existing CrewAI test is unrelated wiring coverage.	Add a regression test for SequentialTaskAgentComponent.build_agent_and_task asserting Agent llm/tools are converted, and verify the SequentialCrew path still passes.
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.
Test Quality And Coverage	⚠️ Warning	No test exercises SequentialTaskAgentComponent.build_agent_and_task; existing CrewAI tests only cover wiring/LLM factory, so the new convert_llm/convert_tools path is unverified.	Add unit tests mocking crewai to assert Agent gets converted llm/tools in SequentialTaskAgentComponent, plus a SequentialCrew regression test that consumes the output.

✅ Passed checks (6 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly describes the main fix in SequentialTaskAgentComponent and matches the changeset.
Linked Issues check	✅ Passed	The PR implements the issue's requested eager conversion of LLMs and tools in SequentialTaskAgentComponent.
Out of Scope Changes check	✅ Passed	The changes are tightly scoped to the CrewAI conversion fix and do not add unrelated behavior.
Test File Naming And Structure	✅ Passed	PASS: The PR only changes a non-test component; no test files were added/modified. Existing tests use the expected pytest naming, class-based structure, and descriptive test names.
Excessive Mock Usage Warning	✅ Passed	No PR-related test file shows excessive mocking; the only relevant LFX test file has no mock usage.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@src/lfx/src/lfx/components/crewai/sequential_task_agent.py`:
- Around line 121-124: The SequentialCrew flow is double-converting tools, which
causes failures when `SequentialTaskAgent` passes already-converted CrewAI tools
through `convert_tools` and `BaseCrewComponent.get_tasks_and_agents` converts
them again. Update `convert_tools` in the crew helper to be idempotent by
checking whether each item is already a CrewAI `Tool` before calling
`Tool.from_langchain`, or otherwise skip re-conversion in the
`get_tasks_and_agents` path when `agent.tools` has already been converted. Focus
on the `SequentialTaskAgent` setup and the shared `convert_tools` utility so
repeated conversion no longer breaks execution.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 20656c1a-1992-4bee-bb37-0aed1a1d1b20

📥 Commits

Reviewing files that changed from the base of the PR and between a277209 and fcaae6b.

📒 Files selected for processing (1)

• src/lfx/src/lfx/components/crewai/sequential_task_agent.py

[coderabbitai]

🩺 Stability & Availability | 🟠 Major

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
# Inspect convert_tools / get_tasks_and_agents and any idempotency handling
fd -t f 'crew.py' src/lfx/src/lfx/base/agents/crewai --exec sed -n '100,175p' {}

# Look for how SequentialCrewComponent feeds agents into get_tasks_and_agents
rg -nP -C3 'get_tasks_and_agents|convert_tools' src/lfx/src/lfx/components/crewai

Repository: langflow-ai/langflow

Length of output: 7510

---

🌐 Web query:

crewai Tool.from_langchain behavior when passed an existing crewai Tool object

💡 Result:

The crewAI Tool.from_langchain method is designed to convert LangChain tools into crewAI-compatible tools [1][2]. When passed an object that is already a crewAI Tool (specifically a subclass of BaseTool), the behavior is generally to attempt a conversion that may result in redundancy, type conflicts, or Pydantic validation errors [3][4]. Because crewAI agents and tasks strictly require tools that inherit from crewAI's BaseTool [5][4], passing an object through Tool.from_langchain that is already structured as a crewAI tool often causes issues during initialization [3]. Users frequently encounter Pydantic validation errors (e.g., 'Input should be a valid dictionary or instance of BaseTool') because the tool is already in the expected format, or the conversion wrapper creates a mismatch between expected schemas [3][5][4]. If you have an existing crewAI tool, you should not pass it to Tool.from_langchain. Instead, you should pass the instance directly to the Agent's tools list [1][6]. If you have a native LangChain tool that needs to be used as a crewAI tool, the recommended practice is to either use the tool decorator or subclass BaseTool to manually wrap the LangChain logic, ensuring compatibility with the current crewAI framework [7][4].

Citations:

• 1: https://docs.crewai.com/en/tools/ai-ml/langchaintool
• 2: https://docs.crewai.com/v1.14.7/en/tools/ai-ml/langchaintool
• 3: [BUG] Agent initialization with native langchain tools produce pydantic validation error crewAIInc/crewAI#1676
• 4: https://stackoverflow.com/questions/79226122/crewai-error-when-trying-to-use-the-langchain-tool-input-should-be-a-valid-dict
• 5: https://community.crewai.com/t/passing-crewstructuredtool-to-tools/4102
• 6: https://docs.crewai.com/v1.14.7/en/tools/overview
• 7: https://docs.crewai.com/en/learn/create-custom-tools

---

Double conversion of tools causes runtime failure in SequentialCrew flow

The convert_tools function in src/lfx/src/lfx/base/agents/crewai/crew.py lacks an idempotency guard and unconditionally applies Tool.from_langchain. In the SequentialCrew path, SequentialTaskAgent initializes agents with already-converted CrewAI tools via tools=convert_tools(self.tools). These agents are then passed to get_tasks_and_agents, which calls convert_tools again on agent.tools. Since Tool.from_langchain expects LangChain tools and fails on existing CrewAI instances (raising Pydantic validation errors or type mismatches), this double conversion breaks the execution.

Double conversion evidence

1. Initial conversion (src/lfx/src/lfx/components/crewai/sequential_task_agent.py:124):

   tools=convert_tools(self.tools),

2. Re-conversion (src/lfx/src/lfx/components/crewai/sequential_crew.py:30 → src/lfx/src/lfx/base/agents/crewai/crew.py:156):

   # In BaseCrewComponent.get_tasks_and_agents
   for agent in agents_list:
       agent.tools = convert_tools(agent.tools)  # Fails on already-converted tools

Add an isinstance(tool, Tool) check in convert_tools to skip already-converted objects, or ensure agents bypass re-conversion if tools are pre-converted.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/lfx/src/lfx/components/crewai/sequential_task_agent.py` around lines 121
- 124, The SequentialCrew flow is double-converting tools, which causes failures
when `SequentialTaskAgent` passes already-converted CrewAI tools through
`convert_tools` and `BaseCrewComponent.get_tasks_and_agents` converts them
again. Update `convert_tools` in the crew helper to be idempotent by checking
whether each item is already a CrewAI `Tool` before calling
`Tool.from_langchain`, or otherwise skip re-conversion in the
`get_tasks_and_agents` path when `agent.tools` has already been converted. Focus
on the `SequentialTaskAgent` setup and the shared `convert_tools` utility so
repeated conversion no longer breaks execution.