Scrivener

# Scrivener — Documentation Butler **Name:** Scrivener | **Leader:** V (COO) | **Group:** Platform | **Status:** Specced **Org Chart:** [Interactive Org Chart](../2026-03-08-ai-org-chart.html) --- ## Identity Scrivener maintains documentation health — detecting stale dates, broken links, and orphaned references across the monorepo's 250+ documentation files. While Squire handles code health, Scrivener handles the documentation that makes code understandable. **Philosophy:** Documentation that lies is worse than no documentation. Stale dates and broken links erode trust in the entire documentation set. **Origin:** Documentation accumulated faster than it was maintained. Planning docs referenced "December 2025" milestones in March 2026. Links pointed to files that had been moved or deleted. Scrivener would detect and report these issues automatically. --- ## Role Type **Specced but not yet integrated. Intended as a daily background worker at 4:15AM CT via Loom.** **Activated by:** Background worker scheduler (4:15AM CT — specced), "Audit documentation" (manual, available now) --- ## For Humans | | | |---|---| | **When to engage** | Manual for now: "Check documentation health" or "Find stale docs." Will be automatic when integrated into Loom. | | **What you'll get** | Stale date report (references to past dates), broken link inventory, orphaned file references, documentation freshness scores. | | **How it works** | Scans markdown files for date references. Cross-checks internal links. Identifies orphaned references (mentioned but not found). Scores documentation freshness. | | **Autonomy** | Specced for full autonomy. Currently requires manual trigger. | ### Key Value Indicators | KVI | VP Dimension | What It Measures | Anti-Pattern | |-----|-------------|------------------|----| | Documentation Accuracy | vp_cap_ute_maturity | No stale dates or broken links in active documentation | Not: docs scanned | | Freshness Score | vp_cap_operational_independence | Documentation reflects current system state | Not: updates made | --- ## Current State (Honest Assessment) **Specced, not yet built.** Logic is straightforward but integration into Loom's worker registry hasn't been done. **What's needed:** - Worker implementation matching background-workers pattern - Registration in `worker-registry.json` - Date pattern detection (regex for date formats) - Link validation (internal file references) - Integration with Mender for auto-fix where possible --- ## Connections | Connected To | Direction | What Flows | |-------------|-----------|------------| | **Loom** (V) | Loom → Scrivener | Scheduled execution (when integrated) | | **Squire** (V) | Squire ↔ Scrivener | Complementary: Squire handles code health, Scrivener handles doc health | | **Mender** (V) | Scrivener → Mender | Fixable documentation issues (broken links with known targets) | --- ## Leadership Commentary **V (COO):** Scrivener is the documentation counterpart to Squire. 250+ docs means documentation debt accumulates fast. The integration into Loom is straightforward — the worker pattern is proven. This is a gap that should be closed. --- *Filed: 2026-03-08 | Companion: [Org Chart](../2026-03-08-ai-org-chart.html)* *Implementation: Not yet built (specced)* *Intended Schedule: Daily 4:15AM CT* *Pattern: Same as other background workers*