test(tools): comprehensive test suite for code_execute module (+45 tests)

[manus-use]

Summary

Adds a comprehensive test suite for the code_execute tool module (src/manus_agent/tools/code_execute.py), which previously had zero test coverage.

What's Tested (45 tests)

CodeExecutor class

• Initialization with default config (Config.from_file()) and explicit config
• Sandbox lazy initialization (_get_sandbox): disabled returns None, enabled creates/starts DockerSandbox, reuses instance on subsequent calls

execute_python

• Sandbox path: routes to sandbox.execute_code() with correct args
• Local path: writes temp file, runs via subprocess.run, cleans up temp file
• Custom timeout passthrough
• Non-zero exit code handling

execute_bash

• Sandbox path: routes to sandbox.execute_command()
• Local path: uses subprocess.run with shell=True
• Failure handling (non-zero exit codes)

cleanup

• Stops running sandbox and resets state
• No-op when no sandbox was created

get_executor (global singleton)

• Creates and reuses a single instance
• Config passthrough

code_execute (async function)

• Language routing: python, bash, sh, shell
• Case-insensitive language matching
• Unsupported language returns error dict
• Non-zero exit code sets error field
• Zero exit code leaves error as None
• asyncio.TimeoutError handling
• Generic exception handling (RuntimeError, PermissionError, OSError)
• Custom timeout passthrough

code_execute_sync (@tool decorator)

• Wraps async function in a new event loop
• Passes all arguments through

Integration tests (subprocess-mocked)

• Full Python execution flow
• Full bash execution flow
• subprocess.TimeoutExpired surfaces as execution error
• Syntax error handling

Edge cases

• Empty code, multiline code, large output
• Both stdout and stderr captured simultaneously

No Overlap with Existing PRs

Checked all 50 open PRs (#80–#140) and 30 most recently merged PRs. Test suites exist for:

• test(core-tools): comprehensive test suite for check_cisa_kev, get_nvd_data, get_cwe_details, get_otx_cve_details #89: check_cisa_kev, get_nvd_data, get_cwe_details, get_otx_cve_details
• test(tools): comprehensive test suite for search_exploit_db, search_packetstorm, query_threat_intelligence_feeds, obtain_cves #90: search_exploit_db, search_packetstorm, query_threat_intelligence_feeds, obtain_cves
• test(tools): comprehensive test suite for submit_cves, tool_output_logger, and verify_exploit helpers #93: submit_cves, tool_output_logger, verify_exploit
• test(cli): comprehensive test suite for manus-agent osv subcommand + README docs #120: osv CLI tests
• test(cli): comprehensive test suite for manus-agent blast-radius subcommand #121: blast-radius CLI tests
• tests: comprehensive CLI tests for _run_analyze, _run_remediate, _run_discover (+78 tests) #123: analyze/remediate/discover CLI tests
• test(cli): comprehensive test suite for manus-agent history subcommand (+39 tests) #132: history CLI tests
• test(tools): comprehensive test suite for track_vendor_response (+67 tests) #137: track_vendor_response tests
• test(cli): comprehensive test suite for manus-agent poc-search subcommand (+57 tests) #139: poc-search CLI tests
• test(tools): comprehensive test suite for http_request truncation wrapper (+36 tests) #140: http_request tests

None test code_execute.py. This module is the core code execution engine (async CodeExecutor with sandbox/local paths, language dispatch, timeout handling, global singleton) and had zero coverage.

Test Results

1203 passed, 3 deselected, 3 warnings in 24.47s

Baseline: 1158 → New total: 1203 (+45 tests), 0 failures.

Technical Notes

• All tests are 100% mocked — no real subprocess calls, no Docker dependency
• Uses pytest-asyncio for async test methods
• Global _executor singleton is reset between tests via autouse fixture
• Mocks DockerSandbox at the import boundary to test sandbox path without Docker