fix: workflow incorrectly marked as completed while nodes are still executing

[tomerqodo]

Benchmark PR from qodo-benchmark#441

[sentry]

Bug: The reordering of propagate_skip_from_edge before mark_edge_skipped causes propagation to halt prematurely because the edge's state is still UNKNOWN when checked downstream.
_{Severity: HIGH}

Suggested Fix

Revert the order of the calls within the for loop in _propagate_skip_to_node. Call self._state_manager.mark_edge_skipped(edge.id) before calling self.propagate_skip_from_edge(edge.id) to ensure the edge's state is updated before propagation continues.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent.
Verify if this is a real issue. If it is, propose a fix; if not, explain why it's not
valid.

Location: api/core/workflow/graph_engine/graph_traversal/skip_propagator.py#L82-L86

Potential issue: In `_propagate_skip_to_node`, the call to `propagate_skip_from_edge`
now occurs before `_state_manager.mark_edge_skipped`. When `propagate_skip_from_edge` is
called for an edge, it analyzes the states of incoming edges for the downstream node.
Because the current edge has not yet been marked as skipped, its state is `UNKNOWN`.
This causes the `analyze_edge_states` logic to return early, halting the skip
propagation prematurely. As a result, downstream nodes that should be skipped will
remain in an `UNKNOWN` state, potentially causing the workflow to be marked as complete
incorrectly.

_{Did we get this right? 👍 / 👎 to inform future reviews.}