wise-workflow-resumelisted

# /wise-workflow-resume — resume an interrupted run ## Why this skill exists A run can stop mid-execution for several reasons — the user picked `Pause` in a wave-sync prompt, the Claude Code session closed while a step was in flight, or a step failed and the user wants to try again after fixing something out of band. The run directory and `state.yaml` survive all of these. This skill picks the run back up where it left off. It shares almost all of its loop with `wise-workflow-run`; the key differences are the preamble (load state instead of build state, reset in-flight steps) and the fact that pre-flight prompts are **skipped** — control mode and worktree choice were already made and are recorded in `state.yaml`. ## Arguments Read `$ARGUMENTS`. The first whitespace-separated token is the `run-id` ULID. When `$ARGUMENTS` is empty, [§2](#2-resolve-the-run-id) prompts the user to pick from the workspace's non-terminal runs. - `run-id` — ULID of the run under `$RUNS_ROOT/<run-id>/` (where `$RUNS_ROOT` is `$(python3 .../workflows.py runs-root)` — resolves to `~/.local/share/wise/runs/<cwd-slug>/` by default). When absent, [§2](#2-resolve-the-run-id) prompts the user to pick from the workspace's non-terminal runs. If `$ARGUMENTS` is empty and no resumable runs exist in the workspace, stop with an error pointing at `/wise-workflow-status` to list runs on disk. ## Procedure ### 1. Init-check + resolve — in ONE message Run the init-check per `${CLAUDE_PLUGIN_ROOT}