Wiki Contradiction and Staleness Audit
Find contradictions, superseded and stale claims, and uncited statements in a wiki, with quoted evidence and a fix for each.
by AgentVoltv1.0.0
Free to sign up · every skill included with AgentVolt Pro
Before & after
Same prompt, same agent — the only difference is the skill.
Prompt“Lint our internal wiki for outdated docs, broken links, and formatting issues.”
Beforegeneric AI agent
A wiki linter checks documentation for consistency, broken references, and staleness. You'd typically look for missing links, outdated information, inconsistent formatting, and pages that haven't been updated in a long time. I'd recommend establishing what 'outdated' means first—is something stale after 6 months? Do you have pages that require regular review cycles? Should you automate this check or do it manually? What are your main pain points with the wiki right now?
Afterwith Wiki Contradiction and Staleness Audit
INTERNAL WIKI LINTING REPORT EXECUTIVE SUMMARY → Total pages scanned: 247 → Critical issues: 18 (block reader understanding) → Warning issues: 34 (inconsistency, minor friction) → Green issues: 195 (compliant) → Recommendation: Priority fix 18 critical by EOD CRITICAL ISSUES (Must fix) CATEGORY: Broken Links (8 found) → Page 'API Documentation' links to deleted 'v2-endpoints' page → Page 'Deployment Guide' references dead imgur URL (4 links) → Page 'Troubleshooting' references old Slack channel #dev-help (archived 2026-03) → ACTION: Redirect v2-endpoints to v3-endpoints, fix Slack reference, update image URLs CATEGORY: Dangerously Outdated (6 found) → 'Getting Started' references Node 14 (current: 18+) → 'Database Setup' mentions Postgres 11 (current: 15+) → 'Production Checklist' references deprecated env var MONGODB_URI (removed 2026-04) → 'Release Process' mentions Jenkins (migrated to GitHub Actions 2026-05) → Last updated: 8+ months ago, no reviewer signature → ACTION: Flag for mandatory review by engineering lead, add update date to header CATEGORY: Missing Critical Info (4 found) → 'First Week On-Call' missing link to runbook repository → 'Billing FAQ' never updated after pricing change (2026-02) → 'Performance Baseline' missing thresholds for alerts → ACTION: Add reviewers to CODEOWNERS, set update frequency WARNING ISSUES (Fix within 1 week) CATEGORY: Inconsistent Formatting (12 found) → 'Setup Guide' uses backticks for code; 'Deployment Guide' uses code blocks → Headings inconsistent: some Title Case, some lowercase → Bullet vs numbered lists vary by section → ACTION: Standardize on code blocks, Title Case headers, bullets for non-sequential CATEGORY: Orphaned Pages (8 found) → 8 pages unreachable from main navigation or index → Likely outdated or superseded → ACTION: Add breadcrumb to each or consider archive CATEGORY: Missing Table of Contents (10 found) → Pages over 3 sections lack TOC → Reduces scannability → ACTION: Add auto-generated TOC to 10 pages CATEGORY: Accessibility Issues (4 found) → 4 images missing alt text descriptions → Heading hierarchy broken (h3 before h2) in 2 pages → ACTION: Audit and fix heading structure PRIORITIZATION ROADMAP Today: [ ] Fix 8 broken external links [ ] Update 4 dangerously outdated tech references [ ] Add link to runbook in on-call guide Week 1: [ ] Standardize formatting across 12 inconsistent pages [ ] Add TOC to 10 large pages [ ] Fix image alt text Week 2: [ ] Establish update schedule for 6 high-value pages [ ] Archive or restore 8 orphaned pages [ ] Add reviewer name + date to critical pages ONGOING MAINTENANCE → Wiki lint automation: Run monthly via CI/CD → Freshness policy: Pages over 6 months need reviewer sign-off → Broken link check: Weekly automated scan → Format consistency: Pre-commit hook prevents new issues
About this skill
Know whether your knowledge base can be trusted before an agent or a customer relies on it. Every finding quotes both sides with their sources and dates and proposes a resolution you approve.
What it does
- Finds contradictory values and procedures across pages, after normalizing units, currencies, and periods
- Flags claims overturned by newer sources and time-sensitive facts older than your threshold
- Spot-checks that citations actually support the claims they are attached to
- Catches synthesis pages built on sources that have since changed
- Samples large wikis by impact and states exactly what was covered
What you get
A prioritized audit report (high, medium, low, gaps) with quotes, sources, dates, and a resolution for each finding, plus log entries for every approved correction.
Best for
Support teams whose agents give inconsistent answers, and anyone maintaining a research or company wiki that has grown past manual review.
SKILL.md preview
SKILL.md
---
name: wiki-content-audit
description: Audit what a wiki actually says - contradictions between pages, claims superseded by newer sources, time-sensitive facts gone stale, uncited or mis-cited claims, synthesis pages built on outdated sources, and knowledge gaps - and return a prioritized list with quoted evidence and a proposed resolution for each. Use when the user asks "are there contradictions in the wiki", "what's out of date", "can I trust the knowledge base", "our support articles give different answers", "audit the wiki before the agent uses it", or "which pages need a refresh".
---
# Wiki Contradiction and Staleness Audit
… (sign up to view the full skill)Sign up to view, copy, and install the full skill
More ai & agents skills
View all AI & Agents skills →New Skill Scaffolder
Create a new Claude Code skill correctly: right mechanism, valid name and folder, a triggering description, and a smoke test.
AI & Agents
Skill Library Lint and Overlap Report
Lint every skill in a repo at once: frontmatter, name-folder match, weak descriptions, overlapping triggers, broken refs, plus a CI script.
AI & Agents
Self-Verifying Agent Loop: Plan, Check, Retry, Close
Turns a goal into a task plan with machine-run checks, executes with retry caps, escalates, and refuses to close until verified.
AI & Agents
Skill Test Runner (Trigger and A/B)
Test a skill in fresh sessions: ready-to-run prompts, commands, and rubric for trigger recall, false triggers, and blind A/B.
AI & Agents