fix: pass explicit argv to CLI entrypoint

[seonghobae]

Why

The local bootstrap wrapper had to replace sys.argv before importing the CLI entrypoint. That made the authenticated server invocation implicit and prevented a reliable local orchestrator boundary.

Change

• add an optional argv parameter to contextual_orchestrator.__main__.main
• keep normal console invocation unchanged
• route subcommands and argparse through the supplied argument list
• cover explicit argv and assert that the process argv is unchanged

LineageWeave can now call the upstream entrypoint without monkey patching.

Verification

• uv run pytest tests/test_cli_auth.py -q
• 9 passed

---

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: f8ca96ed-be47-42c9-9b1a-0492db3a5f9a

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[seonghobae]

Exact-head review complete at a2a3edc. The branch is a clean stack on PR #789 (base corrected from main to feat/auto-embedding-model-selection); the effective diff is only the explicit argv CLI boundary and its regression. Local proof: 13 tests passed (CLI auth plus API contract), Ruff, compileall, and git diff --check passed. Auto-merge is being armed; protected approval and hosted Checks remain required.

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no bugs or issues to report.

[seonghobae]

Stack reconciliation completed at exact HEAD 365442b (verify live SHA). Current PR #789 commits are now integrated without dropping its auto-discovery startup behavior; explicit argv is preserved and the effective stack diff remains the CLI boundary plus its regression. Local proof on the reconciled tree: 50 tests passed across CLI/API, discovery, embedding, and security suites; Ruff, compileall, and git diff --check passed. Hosted Checks are re-running on this exact head; protected auto-merge remains enabled and still requires independent approval.

[devin-ai-integration]

Devin Review found 1 new potential issue.

[devin-ai-integration]

🟡 Startup model discovery runs twice, doubling provider queries

When automatic model discovery is requested, the discovery routine is invoked twice back-to-back (_auto_discover_runtime_agents(orchestrator) at contextual_orchestrator/__main__.py:361-362) instead of once, so the server contacts every provider and re-syncs agents a second time for no reason.
Impact: Every server startup with auto-discovery enabled makes a redundant full round of provider network calls and agent syncing, wasting time and provider quota and doubling the chance of a startup failure.

Duplicated conditional block introduced by the PR

The original code had a single guarded call. This PR added an identical second copy:

    if args.auto_discover_model_agents:
        _auto_discover_runtime_agents(orchestrator)
    if args.auto_discover_model_agents:
        _auto_discover_runtime_agents(orchestrator)

_auto_discover_runtime_agents (contextual_orchestrator/__main__.py:254-262) calls discover_all_models() (provider queries) and orchestrator.sync_discovered_agents(...), both of which now execute twice. This duplication is unrelated to the PR's stated intent (passing explicit argv) and is almost certainly an accidental bad merge/edit.

Suggested change

	if args.auto_discover_model_agents:
	_auto_discover_runtime_agents(orchestrator)
	if args.auto_discover_model_agents:
	_auto_discover_runtime_agents(orchestrator)

---

Was this helpful? React with 👍 or 👎 to provide feedback.

[seonghobae]

Superseded by stacked PR #801, which is based on the current #765 head carrying the required auto-discovery and body-limit options.

[seonghobae]

Review repair at current HEAD 365442be60ec6da311d992ddaa69e2bfe2ab1a8a: removed a duplicate \ startup call and added a regression asserting discovery runs exactly once. Local verification: pytest -q tests/test_cli_auth.py tests/test_auto_discovery_server.py (12 passed), Ruff, compileall, and diff --check passed. Base is fb4691838ea193004a8a375a426ec328c3faf1f8; hosted Checks and protected review remain authoritative.