Feature: Retry Patterns Coverage Boost Additional scenarios that exercise previously uncovered code paths in the retry_patterns module. Background: Given the retry patterns coverage module is imported Scenario: get_retry_decorator returns configured decorator for known category When I request the retry decorator for known category "network" Then the returned decorator should use the network category configuration And calling the decorated function should succeed Scenario: get_retry_decorator returns configured decorator for database category When I request the retry decorator for known category "database" Then the returned decorator should use the database category configuration Scenario: RetryContext.execute raises RuntimeError when function returns None Given I have a retry context named "none-returning" And I have a function that always returns None When I execute the None-returning function with the retry context Then a RuntimeError should be raised with message "Retrying must execute at least once" Scenario: RetryContext.async_execute raises RuntimeError when async function returns None Given I have a retry context named "async-none-returning" And I have an async function that always returns None When I execute the async None-returning function with the retry context Then a RuntimeError should be raised with message "AsyncRetrying must execute at least once" Scenario: Auto-debug retry returns result dict without error key Given I have an async function that returns a dict without an error key When I apply auto-debug retry for dict success Then the auto-debug result should be the success dict Scenario: Auto-debug retry returns non-dict result Given I have an async function that returns a non-dict result When I apply auto-debug retry for non-dict success Then the auto-debug result should be the non-dict value Scenario: Auto-debug retry returns None when all attempts pass without error Given I have an async function that returns None from every attempt When I apply auto-debug retry for None result exhaustion Then the auto-debug result should be None Scenario: log_after_retry success path falls back when logger rejects keywords Given a logger that rejects keyword arguments is installed for coverage boost When I call log_after_retry with a successful outcome Then the fallback success message should be logged Scenario: log_after_retry failed path covers else-None branch in exception ternary Given a logger that accepts all arguments is installed for coverage boost And I have a retry state where outcome.failed toggles between checks When I call log_after_retry with the toggling outcome Then the exception keyword should resolve to None