What problem does it solve? Resetting a recruiting-search pipeline normally risks losing expensive LLM-generated artifacts (facts, verdicts, research results) or leaving stale derived state that corrupts a fresh run. This Skill performs a full pipeclean that scrubs all derived state while preserving every paid artifact, so the entire pipeline can be re-walked with cache hits instead of re-spending on API calls. ## Core Features & Use Cases - Safe dry-run preview: Runs read-only by default, showing exactly which paths would move, their sizes, and the backup destination before any change. - Move-never-delete reset: Relocates merged people data, directory, review decisions, import dirs, cluster indexes, and search indexes to a timestamped backup outside the repo with a JSON restore manifest. - Paid-artifact preservation: Keeps OpenAI synthesis facts, cluster pair verdicts, LinkedIn-judge verdicts, deep-research results, profile caches, and message stores keyed by stable identifiers. - Use Case: After a messy review pass, run $clean-slate to reset the pipeline to a fresh-install state, then re-walk $setup, $import-gmail, and $deep-context with every LLM stage cache-hitting at no additional cost. ## Quick Start Ask the agent to run a clean-slate pipeclean of the Powerpacks pipeline, review the dry-run summary, and confirm before applying.
