Python: Fix Hyperlight CodeAct span parenting

[eavanvalkenburg]

Motivation & Context

CodeAct host-tool calls made from inside execute_code were losing their parent span when Hyperlight hopped onto its worker thread. That made the execution flow look flattened in OTEL backends and hid the nested relationship between execute_code and the host tool call.

Description & Review Guide

• What are the major changes? Preserve the active OpenTelemetry context across Hyperlight's worker-thread boundary before running the sandbox callback, and add a regression test that asserts the host-tool span is a child of execute_code.
• What is the impact of these changes? Hyperlight traces now reflect the real nested execution flow, which makes CodeAct runs easier to inspect and debug in observability tools.
• What do you want reviewers to focus on? The context handoff in the Hyperlight worker-thread path and the span-parenting assertion in the regression test.

Related Issue

Fixes #6704

Contribution Checklist

• The code builds clean without any errors or warnings
• All unit tests pass, and I have added new tests where possible
• The PR follows the Contribution Guidelines
• The PR is linked to an issue and there is no other open PR for this issue (see Related Issue above).
• This is not a breaking change. If it is a breaking change, add the breaking change label (or add "[BREAKING]" to the title prefix, before or after any language prefix) — a workflow keeps the label and title prefix in sync automatically.

[github-actions]

Automated Code Review

Reviewers: 5 | Confidence: 93% | Result: All clear

Reviewed: Correctness, Security Reliability, Test Coverage, Failure Modes, Design Approach

---

Automated review by eavanvalkenburg's agents

[Copilot]

Pull request overview

This PR fixes OpenTelemetry span parenting for CodeAct host-tool calls executed inside execute_code when Hyperlight crosses thread boundaries, so traces reflect the true nested execution flow.

Changes:

• Propagate the active contextvars context into Hyperlight’s worker thread executor path.
• Propagate the active contextvars context into the dedicated thread used to run the async sandbox callback (asyncio.run(...)).
• Add a regression test asserting the host tool span is a child of the execute_code tool span.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
python/packages/hyperlight/agent_framework_hyperlight/_execute_code_tool.py	Copies and replays contextvars context across the Hyperlight worker-thread boundary and the sandbox-callback thread boundary to preserve OTEL parent span context.
python/packages/hyperlight/tests/hyperlight/test_hyperlight_codeact.py	Adds an in-memory OTEL exporter fixture and a regression test asserting correct parent/child span relationship for host tool calls inside execute_code.

[github-actions]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
packages/hyperlight/agent_framework_hyperlight
_execute_code_tool.py	623	93	85%	70, 177, 240, 272, 275, 310–311, 326, 328, 341, 359, 369, 394, 399, 406, 412, 420, 428–430, 432–437, 477, 482, 484, 486, 511–512, 518–519, 524–525, 544–545, 554–555, 590, 621, 627–630, 648–651, 659, 692–693, 700–701, 703, 713–714, 754–755, 762, 808, 865–871, 940, 967, 1037, 1073, 1079–1081, 1110–1114, 1118–1119, 1124, 1141–1145, 1149–1150, 1207–1208
TOTAL	42156	4981	88%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
8304	37 💤	0 ❌	0 🔥	2m 9s ⏱️