code sandboxes

[echarles]

No description provided.

[Copilot]

Pull request overview

This PR updates agent-codemode to use a variant-neutral CodeSandboxClient API (instead of Sandbox/EvalSandbox) and aligns execution + tests with the new execute_code / execute_code_streaming methods.

Changes:

• Migrates executor/toolset/server integration to CodeSandboxClient and renames sandbox plumbing (_sandbox → _sandbox_client, sandbox → sandbox_client).
• Updates streaming execution tests to validate the streaming pathway and infrastructure-failure reporting.
• Refreshes examples/docs and normalizes dependency names/constraints in pyproject.toml.

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tests/test_skill_bindings.py	Updates test sandbox injection to use _sandbox_client and execute_code.
tests/test_executor_streaming.py	Renames sandbox mocks to client-style API and adds coverage for streaming infra failure behavior.
tests/test_executor_async.py	Updates async example/test script to CodeSandboxClient and execute_code.
pyproject.toml	Normalizes dependency names and updates version constraints (notably MCP and pydantic-ai optional deps).
examples/simple/agent_cli.py	Updates example wiring to create/pass a shared CodeSandboxClient.
docs/docs/skills/index.mdx	Updates documentation snippet to use CodeSandboxClient.create(variant="eval").
agent_codemode/toolset.py	Renames toolset configuration field to sandbox_client and passes it into the executor.
agent_codemode/server.py	Uses executor.sandbox_client to inject context variables.
agent_codemode/composition/executor.py	Core migration to CodeSandboxClient, adds streaming collection logic, and updates sandbox lifecycle calls.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.