piro-runlisted

# piro-run - subagent-driven execution of a Kiro spec piro's executor. Reads the 3 files in `.kiro/specs/<slug>/` (requirements.md / design.md / tasks.md), has a fresh subagent implement the tasks in tasks.md one at a time, puts each through a review for spec conformance and quality, and marks passing tasks `[x]` before moving on. The Kiro-spec adaptation of superpowers:subagent-driven-development. Generating and changing specs is piro's own job (this skill does not rewrite specs. The one exception is the checkboxes in tasks.md). ## Prerequisites - The 3 files in `.kiro/specs/<slug>/` must be present in the target project - Run on a git working branch. **Running directly on main/master is forbidden** (unless the user explicitly agrees). If there are uncommitted changes, stash them first ## Flow ### 0. Reading the spec and planning execution 1. If no slug is given, list what's under `.kiro/specs/` and confirm 2. Read requirements.md / design.md / tasks.md 3. Check the state of tasks.md: - `[x]` is complete. **Never re-run it** (the checkboxes in tasks.md are the source of truth for progress) - `[ ]` is what you execute, in the order written - `[ ]*` (optional) is also executed by default. Skip it only when the user says something like "skip the tests", leave it as `[ ]`, and state it plainly in the completion report 4. If there are contradictions between tasks, or a `_Requirements:_` reference that doesn't exist in requirements.md, raise them with the