pragman-unfcklisted

# Pragman Unfck Turn selected session evidence and the user's missing context into reversible workflow improvements. Diagnose before tuning: logs show observable behavior, not complete intent or causality. ## Use this skill when - Repeated agent sessions show waiting, retries, compaction, scope drift, provider mismatch, or avoidable tool friction. - You can select bounded Codex, Claude Code, or Cursor evidence instead of handing over an entire history. - A proposed workflow change needs a sanitized diff, evaluation, separate apply approval, and rollback point. ## Acceleration payoff It improves the system around the agent—not just one prompt—while preventing noisy logs or private transcripts from becoming unreviewed automation. ## Workflow 1. Find `pragman` through executable lookup and verify `pragman --version --json` against [COMPATIBILITY.md](./COMPATIBILITY.md). Never install it automatically. Without a compatible CLI, disclose that deterministic discovery, parsing, redaction, diffs, evaluation, and rollback are unavailable; continue only with an interview and excerpts the user explicitly supplies. 2. Ask the user to choose source adapters, project aliases, inclusive UTC date range, content categories, and privacy depth. Default to `metadata-only`. Do not silently scan every Codex, Claude Code, or Cursor history. Read [references/privacy-and-ingestion.md](references/privacy-and-ingestion.md) before selecting sources or using excerpts. 3. Preview roots, formats, co