What problem does it solve? Long-running or unattended agent work leaves reviewers unable to reconstruct what was decided, why, and on what evidence without rereading entire transcripts. This Skill keeps a single canonical decision trail so a human can audit the run after the fact. ## Core Features & Use Cases - Structured TSV decision log: One row per decision with timestamp, phase, decision, rationale, evidence pointer, and result, rendered as a sortable table on GitHub or in a terminal. - Safe logging helper: The scripts/log.sh script appends well-formed rows, strips stray tabs and newlines, and prefixes spreadsheet-formula characters to prevent formula injection from generated text. - End-of-run auditing: Walks the log against the run transcript to remove invented rows, verify evidence resolves, and add missing pivots, then requires a cross-model subagent review that flags risks in an "Attention" section. - Use Case: During a multi-week migration, an agent logs each fork, revert, and verification result to decisions.tsv; a reviewer later opens the committed file in the PR and follows each evidence link to confirm the work. ## Quick Start Use the show-me-your-work skill to keep a TSV decision log for this multi-phase refactoring run and audit it against the transcript before handing back.
