What problem does it solve? When an AI session's context is compressed or lost mid-workflow, critical state like worktree paths, base branches, current branches, and step numbers can be forgotten. This Skill solves that by persisting workflow state to per-worktree JSON files, enabling reliable context recovery and parallel multi-session development. ## Core Features & Use Cases - State File Persistence: Writes workflow state (worktree path, base branch, current step, timestamps) to a JSON file inside the git directory, invisible to git status. - Parameterized Workflows: Supports both bug-fix and feature-development workflows via WORKFLOW_TYPE and BRANCH_FIELD parameters. - Recovery & Concurrency Control: Provides a step-by-step recovery procedure when state files are missing, plus concurrency checks that block multiple workflows on the same worktree. - Use Case: After a long bug-fix session gets compacted, the agent reads bug-fix-state.json from the worktree's git dir, restores the branch and step number, and resumes exactly where it left off instead of restarting. ## Quick Start Ask the agent to restore my current workflow context using the shared state file before continuing the bug-fix workflow.