What problem does it solve? Closing a story in an AI-assisted scrum workflow requires manually verifying every task reached a terminal state, digging through git history for actual change metrics, and updating Markdown board files by hand — error-prone busywork that this Skill automates end to end. ## Core Features & Use Cases - Terminal-State Guard: Verifies every task in a story's frontmatter is Passed or Done before allowing closure, halting with a blocking report otherwise. - Actual Diff Metrics: Extracts per-task actual_files_changed and actual_lines_delta from EST-tagged git commits and writes them into task frontmatter, using a full-credit attribution model for bundle commits. - Scope Divergence & Privatized Checks: Computes scope_divergence_flag against configurable thresholds and statically tests each ticket's touched files against .publicignore, marking fully blocklisted tickets as Privatized and logging Publicized/Privatized mismatches to the event log. - Use Case: After the Tester marks all tasks of story E32_S06 as Passed, run the close flow to stamp real diff stats onto every task, flag any task that exceeded its declared execution scope, and set the story status to Done with a completion date. ## Quick Start Ask the agent to close story E32_S06 so it verifies task states, records diff statistics, and marks the story Done on the board.
