progressive-chain-validationlisted

# Progressive Chain Validation Use this Test-window skill to turn a long workflow into a source-derived execution plan, then validate or repair that workflow one node at a time. This is a general Test skill. Product-specific adapters, overlays, examples, and repository names belong in the installed workspace or product repository, not in Wakeflow's reusable template. ## Required Inputs Before acting, identify: - target workflow or feature chain; - target project root, external test project, or sandbox copy; - whether the request is plan-only, plan-then-execute, repair, review, or metrics comparison; - existing requirement design, test card, failing output, benchmark reference, trace, report, or user scenario; - allowed command scope, write boundary, and destructive-operation boundary; - entry hints, relevant documents, existing tests, observability surfaces, and runtime data locations; - desired useful units, baseline references, comparison fixtures, scorecard fields, or metric gates when metrics are in scope; - executor scope such as CLI, API route, MCP handler, dashboard, internal agent, external agent, worker, service, or mixed; - current node or round when resuming; - per-node isolation constraints: simulated data, frozen upstream artifacts, downstream cut points, async controls, and reset requirements; - terminal stability constraints: timeout budget, sync versus async mode, output capture path, exit evidence, and hang recovery rule. If any command ma