ci: install QuarkAgent extras from setup.py, bump langchain to 1.x

[haeter525]

Description

pytest.yml hardcoded langchain/langchain-core/langchain-openai versions in a standalone pip install step, completely independent of setup.py's quarkAgentRequirements. Any dependabot bump to those three packages in setup.py showed all-green CI while the hardcoded line silently kept testing the old versions — the bump was never installed, let alone tested. This PR makes setup.py the single source of truth and bumps the trio to the latest mutually-compatible release.

Key Changes

• .github/workflows/pytest.yml: drop the hardcoded pip install langchain==... langchain-core==... langchain-openai==... line; install via pip install ".[QuarkAgent]" instead, so CI installs whatever setup.py actually declares.
• setup.py: bump quarkAgentRequirements to langchain==1.3.10, langchain-core==1.4.8, langchain-openai==1.3.2.
• quark/agent/agentTools.py: @tool decorator import moved from langchain.agents to langchain.tools (LangChain 1.0 relocated it).
• quark/agent/quarkAgent.py, quark/agent/quarkAgentWeb.py: migrated off the removed AgentExecutor + LCEL chain pattern to langchain.agents.create_agent. Conversation history keeps the same HumanMessage/AIMessage shape; the reply now comes from response["messages"][-1].content instead of response["output"].
• tests/agent/test_agentTools.py: testColorizeInColor called the @tool-wrapped function directly; StructuredTool is no longer callable that way in 1.x, switched to .func(...) like every other test in the file already does.

Motivation and Context

Issue 18z/QuarkHQ#3 (automating dependency-update PR review) surfaced this as a recurring false-positive: dependabot PRs bumping the langchain stack always showed green CI without the bump ever being exercised. Fixing the CI install path and doing the real version bump removes the need to treat these three packages as a special case going forward.

How Has This Been Tested?

• pip install ".[QuarkAgent]" resolves cleanly to the new trio, no ResolutionImpossible.
• pytest tests/agent/ — 16/16 pass.
• Both quarkAgent.py's real CLI entry point and quarkAgentWeb.py's exact construction code build a real create_agent graph with no network call (verified manually).
• Not verified: the live agent loop itself (a real LLM call) — needs a real OPENAI_API_KEY, not available in this environment. Flagging as a known gap rather than claiming full verification.

Checklist

• Follows project commit/style conventions
• Self-reviewed (ran /code-review: black/pylint/mypy/bandit/codespell, fixed the one real regression found — a type annotation gap on conversationHistory that only mattered because the new create_agent API is strictly typed)
• Documentation — N/A, no public quark.script API changed
• No new warnings beyond known pre-existing ones (mypy --strict has pre-existing errors across this module unrelated to this change; not in scope)
• Existing unit tests pass locally (tests/agent/, 16/16)
• Live agent loop tested end-to-end with a real API key — not possible in this environment

[codecov]

Codecov Report

❌ Patch coverage is 61.53846% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 81.21%. Comparing base (f90ae20) to head (26a0626).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
quark/agent/quarkAgentWeb.py	0.00%	4 Missing ⚠️
quark/agent/quarkAgent.py	80.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #930   +/-   ##
=======================================
  Coverage   81.21%   81.21%           
=======================================
  Files          81       81           
  Lines        7122     7122           
=======================================
  Hits         5784     5784           
  Misses       1338     1338           

Flag	Coverage Δ
unittests	81.21% <61.53%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.