Skip to content

vt-v-audit

Audit VMS vault content quality. Validates YAML frontmatter compliance, checks cross-reference integrity, scores content depth, detects skeleton content, and calculates compliance-score. Produces a structured audit report. Use after content development batches or for periodic quality assessment.

Plugin: vms
Category: Other
Command: /vt-v-audit

VMS Content Audit

This skill validates vault content quality across multiple dimensions and produces a structured audit report.

Process

  1. Determine scope from argument:
  2. --full — Audit entire vault (all domains)
  3. --changed — Audit only git-changed files (git diff --name-only)
  4. --cwp CWP-NN — Audit only files belonging to a specific CWP (resolved via cwp_file_resolver.py)
  5. --snapshot — Output machine-readable JSON metrics (combine with --cwp or --full)
  6. <path> — Audit specific path or glob
  7. Run audit scripts from scripts/ directory:
  8. yaml_audit.py — YAML schema validation
  9. cross_ref_checker.py — Wiki-link integrity (forward links)
  10. impact_analyzer.py — Reverse reference integrity (inbound links)
  11. reifegrad_validator.py — Reifegrad cross-validation across shared norm_refs
  12. field_consistency_checker.py — Review date and approved_by consistency across shared norm_refs

4b. Reifegrad Cross-Validation

  • Run reifegrad_validator.py to detect maturity level inconsistencies
  • For each norm_ref, collects all control files and their reifegrad values
  • Flags when reifegrad spread exceeds threshold (default: >1 level difference)
  • Skips reifegrad: na entries and non-control file types
  • Output includes affected files, min/max reifegrad, and assessment dates

5. Reverse Reference Integrity

  • Run impact_analyzer.py in summary mode to build the full reverse index
  • Top-10 most-referenced files (highest inbound [[wikilink]] count)
  • Stale references: files whose review_date is older than the last edit of files that reference them
  • Orphaned files: zero inbound references, excluding index.md and dashboard pages
  • HIGH priority flag: files with 10+ inbound references and review_date older than 90 days

5b. Change Cascade Report

  • Run cascade_report.py to trace impact of git changes through the reverse wikilink index
  • Accepts --ref-range (default: main..HEAD) or explicit --files list
  • Produces tree-format report: changed file → directly affected → transitively affected
  • Respects --max-depth (default: 3) to cap cascade traversal
  • Domains (ISMS, QMS, DSMS, VMS) shown per affected file for cross-domain awareness

5c. Field Consistency Check

  • Run field_consistency_checker.py to detect review_date and approved_by inconsistencies
  • For each norm_ref, collects all files and their review_date and approved_by values
  • Flags review_date spread > 6 months (configurable via --threshold-months)
  • Reports approved_by mismatches within same norm domain (informational, not error)
  • Output includes affected files, date spread, and approver discrepancies

6. Compliance Score

  • score = (populated_fields / expected_fields) * 100
  • Per-file, per-domain, and vault-wide

Report Location

Reports are written to: reports/vt-v-audit-YYYY-MM-DD.md

Read references/audit-criteria.md for detailed pass/warn/fail thresholds.

Integration

  • Run after each /vt-v-write batch to validate content
  • Run before /vt-v-sprint to establish baseline
  • Results feed into the next /vt-v-plan sprint planning
  • --snapshot mode is used by /vt-v-cwp to record metrics before/after CWP work
  • CWP file resolution uses cwp_file_resolver.py from /vt-v-cwp skill

Open Brain Capture (Optional)

After completing the audit, if the capture_thought MCP tool is available, capture the audit outcome.

How: 1. Check if capture_thought tool is available. If not: skip silently. 2. Call capture_thought with: 

thought: "VMS audit {date}: compliance-score {X%}. Files audited: {N}. Findings: {critical} critical, {warning} warning, {info} info. Top issue: {description}. Management system: {ISMS/DSMS/QMS}."
3. On timeout or error: log debug message and continue. Never fail the skill.