fix(workflows): engine loop cap ignores bool max_iterations

[jawwad-ali]

Description

The workflow engine's while/do-while loop-cap guard:

max_iters = step_config.get("max_iterations")
if not isinstance(max_iters, int) or max_iters < 1:
    max_iters = 10
...
for _loop_iter in range(max_iters - 1):

does not fall back to the default for a boolean max_iterations. Since isinstance(True, int) is True and True < 1 is False, a max_iterations: true slips past the guard and the loop runs range(True - 1) == range(0) — capping at a single iteration instead of the default 10.

This is the same bool-is-int trap already fixed in the while/do-while validators (#3237); the engine's runtime path had the same gap, and WorkflowEngine.execute() does not auto-validate, so this guard is the only line of defence at run time. The function's sibling continue_on_error is True handling already special-cases bools.

Fix

if (
    isinstance(max_iters, bool)
    or not isinstance(max_iters, int)
    or max_iters < 1
):
    max_iters = 10

Testing

• New test_loop_with_bool_max_iterations_uses_default_cap: an unvalidated from_string workflow with max_iterations: true and a condition that stays truthy; asserts the tick counter reaches 10 (default cap). Fails before ('1' == '10' — capped at one iteration, verified by source-stash), passes after.
• uvx ruff check clean; the existing test_while_loop_runs_to_max_when_condition_stays_true (int max_iterations) stays green.

AI Disclosure

• I did use AI assistance (describe below)

Found and fixed with Claude Code (Claude Opus 4.8) under my direction. AI flagged the runtime gap left by the validator-only #3237 fix; I confirmed the single-iteration cap, proved fail-before via source-stash, and reviewed the diff before submitting.

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.