fix(tools): raise ToolUsageError instead of bare raise for non-dict arguments

[ashusnapx]

Summary

Fixes RuntimeError: No active exception to re-raise in ToolUsage._original_tool_calling() when _validate_tool_input returns a non-dict and raise_error=True.

Closes #6430

Problem

_original_tool_calling() at tool_usage.py:853 had a bare raise outside any except block:

if not isinstance(arguments, dict):
    if raise_error:
        raise   # <-- no active exception here
    return ToolUsageError(...)

When this path was hit with raise_error=True, Python raised RuntimeError: No active exception to re-raise instead of the intended ToolUsageError. The _tool_calling() fallback relies on catching a meaningful exception to trigger its retry logic, so this produced a confusing RuntimeError.

Also flagged by ruff check --select PLE0704.

Fix

Replace bare raise with explicit raise ToolUsageError(...):

if not isinstance(arguments, dict):
    if raise_error:
        raise ToolUsageError(
            f"{I18N_DEFAULT.errors('tool_arguments_error')}"
        )
    return ToolUsageError(f"{I18N_DEFAULT.errors('tool_arguments_error')}")

The except block above (line 846-848) correctly uses bare raise inside an active exception context — left unchanged.

Tests (4 new)

Test	What it verifies
test_original_tool_calling_non_dict_raises_tool_usage_error	raise_error=True raises ToolUsageError, not RuntimeError
test_original_tool_calling_non_dict_returns_error	raise_error=False returns ToolUsageError
test_tool_calling_fallback_on_non_dict	Full _tool_calling flow falls back gracefully
test_non_dict_raises_not_runtime_error	Exception type is ToolUsageError, not RuntimeError

All 32 tests in test_tool_usage.py pass. ruff check --select PLE0704 clean.

Verification

pytest lib/crewai/tests/tools/test_tool_usage.py -v
ruff check lib/crewai/src/crewai/tools/tool_usage.py --select PLE0704

Summary by CodeRabbit

• Bug Fixes

  • Improved error handling when tool input can’t be parsed into valid arguments, so users now get a clear tool usage error instead of an unexpected crash.
  • Preserved fallback behavior so invalid tool input is reported consistently rather than stopping execution abruptly.

• Tests

  • Added regression coverage for invalid tool input cases to verify the correct error type and message are returned in both strict and non-strict flows.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 21077f8a-20f4-4527-867b-2ceeda2fe014

📥 Commits

Reviewing files that changed from the base of the PR and between 289686a and a395ea2.

📒 Files selected for processing (2)

• lib/crewai/src/crewai/tools/tool_usage.py
• lib/crewai/tests/tools/test_tool_usage.py

---

📝 Walkthrough

Walkthrough

The _original_tool_calling method in ToolUsage now raises ToolUsageError with the localized tool_arguments_error message when parsed tool arguments are not a dict and raise_error=True, replacing a bare raise that previously caused a RuntimeError. Regression tests were added covering this behavior.

Changes

Bare Raise Fix

Layer / File(s)	Summary
Raise ToolUsageError for non-dict arguments lib/crewai/src/crewai/tools/tool_usage.py	_original_tool_calling now raises ToolUsageError with the tool_arguments_error i18n message instead of executing a bare raise when parsed arguments are not a dict and raise_error=True.
Regression tests for non-dict input lib/crewai/tests/tools/test_tool_usage.py	Adds _make_tool_usage_for_non_dict_test() helper and four tests verifying _original_tool_calling raises/returns ToolUsageError (not RuntimeError) for non-dict parsed input, and that _tool_calling falls back correctly.

Estimated code review effort: 2 (Simple) | ~10 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 62.50% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main fix: replacing a bare raise with ToolUsageError for non-dict tool arguments.
Linked Issues check	✅ Passed	The code and regression tests address #6430 by raising ToolUsageError, preserving fallback behavior, and covering the non-dict path.
Out of Scope Changes check	✅ Passed	The changes stay within the bug fix scope and add only targeted regression tests, with no unrelated behavior changes evident.

✨ 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.}