fix(agents): guard on chat.messages not chat.last to avoid IndexError in _run_tools

[chuenchen309]

What

ChatWorkflow(...).run() with no messages added crashes with IndexError instead of the intended no-op.

The "no tool calls" guard in _run_tools (libs/giskard-agents/src/giskard/agents/workflow.py) starts with:

if not chat.last or ...:
    return

But Chat.last is self.messages[-1], which raises IndexError: list index out of range on an empty messages list rather than returning something falsy. Since messages defaults to [], a workflow that runs before any .chat() call hits this — the IndexError propagates and run() (default RAISE error policy) wraps it in WorkflowError("Step processing failed") instead of the intended early return.

Fix

Guard on chat.messages (the list) instead of chat.last (which indexes into it):

if not chat.messages or ...:
    return

One-line change in workflow.py.

Testing

Added test_run_with_no_messages_does_not_raise_indexerror to libs/giskard-agents/tests/test_workflow_steps.py (mocked BaseGenerator, no real API calls). TDD red confirmed against the unfixed guard (raises WorkflowError wrapping the IndexError), green after the fix. libs/giskard-agents/tests/test_workflow_steps.py (non-functional): 4 passed. ruff check + ruff format --check clean.

xref: no open PR or issue touches workflow.py, _run_tools, Chat.last, or ChatWorkflow.

---

Disclosure: this PR was drafted with AI assistance (Claude); I reviewed, tested, and take responsibility for the change.

[gemini-code-assist]

Code Review

This pull request fixes a potential IndexError in the _run_tools method of ChatWorkflow by checking if chat.messages is empty instead of checking chat.last (which accesses messages[-1]). It also adds a corresponding unit test to verify this behavior and prevent regressions. There are no review comments to address, and I have no additional feedback to provide.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[kevinmessiaen]

LGTM, thanks for the fix!