verifiers: fix tool call rendering from saved outputs

[snimu]

Description

Since tools are json-serialized, they are no longer correctly displayed in the tui:

This PR fixes that problem:

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Test improvement

Testing

• All existing tests pass when running uv run pytest locally.
• New tests have been added to cover the changes

Checklist

• My code follows the style guidelines of this project as outlined in AGENTS.md
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• Any dependent changes have been merged and published

---

Note

Low Risk
Low risk refactor/bugfix limited to Rich formatting utilities and call sites; main risk is minor display regressions or unexpected tool-call string shapes causing JSON decode errors.

Overview
Fixes Rich/TUI rendering of prompts/completions when tool_calls are stored as JSON strings by introducing a shared format_messages() helper that deserializes and pretty-prints tool calls.

Removes duplicated message-formatting code in eval_display.py and logging_utils.py and switches both to the new shared formatter to keep prompt/completion display consistent.

^{Written by Cursor Bugbot for commit c684082. This will update automatically on new commits. Configure here.}

[cursor]

Unhandled JSON parse error in tool call normalization

Medium Severity

The new normalize_tool_call function calls json.loads(tc) when tc is a string, but this call is not wrapped in a try-except block. If a malformed JSON string is encountered (e.g., corrupted data, old format, or encoding issues), a json.JSONDecodeError will be raised and propagate up, potentially crashing the TUI display. This is inconsistent with the defensive handling of json.dumps(args) later in the same function, which is wrapped in a try-except.

 

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}

[mikasenghaas]

lgtm