fix: thinking block content leak in text extraction

[bkrabach]

Summary

Fixes a thinking-block text leak at two locations in amplifier_module_loop_events/__init__.py.

Root Cause

There are two content block model systems in amplifier-core:

• content_models.ThinkingContent → has .text (was the bug hazard)
• message_models.ThinkingBlock → has .thinking (already safe)

The old execute() inline text extraction used hasattr(block, "text") at both the main loop path (line ~268) and the max-iterations fallback path (line ~621). Because ThinkingContent has a .text attribute, this guard let thinking text leak into final_response.

Fix

Extracts a shared _extract_text_from_content helper (module-level, following the existing _normalize_tool_call convention) that uses an explicit block.type == "text" check. Replaces both call sites with the helper — eliminating the duplication and the bug in one pass.

Pattern

Same fix as:

• amplifier-module-loop-streaming PR #25 (merge df5c0e1) — introduced _extract_text_from_content pattern
• amplifier-module-loop-basic PR fix: pass provider= kwarg to context_manager.get_messages_for_request #11 (fa67fe2) — companion fix, single location

This module has two locations with the bug, so extracting the shared helper is especially important for DRY compliance.

Tests Added

New file tests/test_thinking_block_leak.py with 8 tests:

Test	Purpose
test_thinking_content_does_not_leak_into_final_response	Primary regression — Location 1 (main loop path)
test_thinking_content_filtered_at_max_iterations_fallback	Primary regression — Location 2 (max-iterations fallback)
test_helper_directly_filters_thinking_content	Direct unit test of the shared helper
test_text_content_passes_through	Regression guard — TextContent still works
test_text_block_passes_through	Regression guard — TextBlock still works
test_thinking_block_does_not_leak_into_final_response	Documents ThinkingBlock was already safe
test_dict_text_block_passes_through	Dict {"type": "text"} blocks pass through
test_dict_thinking_block_is_filtered	Dict {"type": "thinking"} blocks filtered

TDD Verification

RED: 4 tests failed before fix (bug tests + import error for helper not yet existing)
GREEN: All 8 tests pass after fix

Note: test_tool_post_no_modify_uses_original in test_hook_modify.py was already failing on main before this branch — not introduced by this change.