feat: implement shared ExecutionContext for concurrent sieve checks

[Jaydeep869]

Description

This PR fully implements the shared ExecutionContext system (previously outlined in comments within darnit/core/models.py). This allows large, heavy API toolings (like OpenSSF Scorecard) to cache and share outputs safely across multiple Control evaluations rather than being run redundantly.

Includes thread safe locking mechanisms get_or_run_tool() to anticipate concurrent sieve handler executions. The created execution_context is seamlessly propagated through CheckContext and injected natively downstream onto custom Python Handler's HandlerContext models.

Motivation and Context

Currently, each control runs tools independently, bottlenecking performance. With heavy tools like OpenSSF scorecard running per rule, wait times escalate rapidly. Shared execution contexts securely cache outputs for multi control usage and prevent duplicate networking trips simultaneously.

How Has This Been Tested?

1. Added unit tests spanning logic within tests/darnit/core/test_models.py verifying cache hits, cache misses, and ThreadPool concurrent access validations limiting tool execution to precisely 1 trigger safely.
2. Verified property logic dynamically within tests/darnit/sieve/test_orchestrator.py via TestExecutionContextPropagation tests.
3. Updated /docs/HANDLER_AUTHORING.md demonstrating custom Python handler integrations.
4. Validated execution using local pytest suite ensuring all core framework / sieve orchestrator paths passed smoothly without regressions.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update

Checklist:

• My code follows the code style of this project (ruff check & ruff format).
• My change requires a change to the documentation.
• I have updated the documentation (docs/HANDLER_AUTHORING.md).
• I have read the CONTRIBUTING document.
• I have added tests to cover my changes.
• All new and existing tests passed.

[mlieberman85]

LGTM — merging. The double-checked locking design is clean, and the 10-thread concurrency test validates the single-execution guarantee correctly.

Three follow-up issues created:

• Wire first heavy tool (e.g., OpenSSF Scorecard) to ExecutionContext.get_or_run_tool #194 — Wire first heavy tool (Scorecard/Inspector/GitHub API) to ExecutionContext.get_or_run_tool so the infrastructure has a real consumer
• Tighten execution_context typing from Any to proper forward reference #195 — Tighten HandlerContext.execution_context typing from Any | None to a proper forward reference
• Document ExecutionContext locking strategy and thread-safety guarantees #196 — Document the locking strategy, concurrency guarantees, and failure semantics

Thanks for turning the long-standing scaffold in core/models.py into a real implementation with good test coverage.