# Freshness Auditor > Run freshness audits and generate prioritized refresh lists for content maintenance - Skill: `uitbreidenos/freshness-auditor` (Agent Skill) - Install (CLI): `npx skillmds@latest add uitbreidenos/freshness-auditor` - Raw SKILL.md: https://api.skillmd.com/api/skills/uitbreidenos/freshness-auditor/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: UitbreidenOS (https://skillmd.com/u/uitbreidenos) - Updated: 2026-09-22 - Page: https://skillmd.com/skills/uitbreidenos/freshness-auditor --- # Freshness Auditor ## When to activate - **Before certification renewal** — audit UitKit content to ensure all files are current and defensible - **Before releasing a new version or major update** — verify documentation accuracy - **Quarterly maintenance cycle** — identify stale content that needs refreshing (every 3 months) - **After a major Claude model release** — flag potentially outdated guidance in skills and workflows - **On-demand content audit** — when you want to assess overall content health without waiting for the quarterly sprint ## When NOT to use - For real-time freshness monitoring (use CI checks with `check-freshness.js` instead) - To validate specific external links (use a dedicated link checker tool) - To rewrite content (use `/workflows/freshness-refresh` instead, which spawns review agents) - To enforce style consistency (use linting rules and `code-review` skill) ## Instructions ### Step 1: Run the freshness audit Execute the freshness check script: ```bash node scripts/check-freshness.js ``` This scans all `skills/` and `agents/` directories for files with an `updated:` date in YAML frontmatter. It flags files older than 6 months as stale and files missing a date entirely. Output includes: - Count of fresh, stale, and undated files - List of stale files with their last-updated date - List of files missing an `updated:` field ### Step 2: Generate the detailed report For a more detailed breakdown organized by category and priority: ```bash node scripts/generate-refresh-report.js > FRESHNESS_REPORT.md ``` This produces a Markdown report with: - Fresh vs stale counts by category (skills, agents, guides, workflows) - Files sorted by age (oldest first) - Estimated time to refresh each file - Tier 1/2/3 categorization for prioritization ### Step 3: Categorize by impact Read the freshness report and sort files into tiers: **Tier 1 (Critical — refresh immediately):** - Core debugging and testing skills - Frequently-referenced guides (getting started, conceptual foundations) - Essential agent definitions used in workflows - Any file older than 12 months These files are on critical paths; staleness could mislead users or break workflows. **Tier 2 (Important — refresh within 2 weeks):** - Domain-specific skills (frontend, backend, DevOps, etc.) - Stack-specific guides and workflows - Secondary agent definitions - Files 6–12 months old in active domains These are referenced regularly but less critical than Tier 1. **Tier 3 (Optional — review, may keep as-is):** - Archived or deprecated skills - Rarely-used agent definitions - Historical examples and case studies - Evergreen conceptual guides unlikely to change These may not need updating if content is still accurate. ### Step 4: Assess content accuracy (manual review) For each Tier 1 file, do a quick accuracy check: ```markdown File: [path] Last updated: [date] Age: [N months] Quick accuracy check: - [ ] All command examples still work in current Claude Code - [ ] Tool names and features referenced still exist - [ ] No references to deprecated features or old model versions - [ ] External links (if any) are not 404 - [ ] Code syntax is current (not using obsolete APIs) Status: [Fresh | Minor Updates Needed | Major Rewrite Needed] ``` If >30% of a file's content is inaccurate or outdated, mark it for a major refresh in the next quarterly cycle. ### Step 5: Produce the prioritized refresh list Create a text file with your findings: ```markdown # Freshness Audit Results — [DATE] ## Summary - Total files scanned: X - Fresh (< 6 months): X - Stale (≥ 6 months): X - Missing dates: X ## Tier 1 (Refresh immediately) - [File path] — last updated [date], [N] months old - ... ## Tier 2 (Refresh within 2 weeks) - [File path] — last updated [date], [N] months old - ... ## Tier 3 (Optional review) - [File path] — last updated [date], [N] months old - ... ## Files needing major rewrite - [File path] — [reason: outdated examples, deprecated features, etc.] ## Next steps 1. Assign Tier 1 files to review agents 2. Schedule Tier 2 review for sprint backlog 3. Archive or deprecate Tier 3 files if no longer relevant 4. Run the full /workflows/freshness-refresh to coordinate updates ``` ### Step 6: Decide on next action Based on the audit results: - **If <10% stale:** Content is healthy. Refresh normally on the quarterly cycle. - **If 10–30% stale:** Schedule a focused refresh sprint for Tier 1 and 2 files within the next 2 weeks. - **If >30% stale:** Critical health issue. Run `/workflows/freshness-refresh` immediately to get all agents updating content in parallel. - **If any files are >12 months old:** Escalate for immediate review. --- ## Example ### Scenario: Pre-certification audit Before certifying UitKit as a production tool, you want to ensure all content is current. ### Execution: ```bash # Step 1: Run freshness check node scripts/check-freshness.js # Output: # Freshness check: 847 files scanned (stale threshold: 6 months) # Fresh: 621 # Stale: 156 # No date: 70 ``` 156 stale files and 70 missing dates — time to remediate before certification. ```bash # Step 2: Generate detailed report node scripts/generate-refresh-report.js > FRESHNESS_REPORT.md # Review the report and categorize: # Tier 1 (Core skills): 32 files # Tier 2 (Domain skills, workflows): 89 files # Tier 3 (Archived, examples): 35 files ``` ### Decision: Based on the audit, the team decides: - Tier 1 files (32) must be refreshed before certification sign-off — assign to 4 agents, 2 hours - Tier 2 files (89) can be batched into 2-week sprints — 3 agents working in rotation - Tier 3 files (35) are archived/deprecated — mark with "archived" tag in frontmatter, no refresh needed ### Next step: Spawn a targeted refresh workflow for Tier 1 files only: ```bash # (Uses the freshness-refresh workflow with --tier 1 flag) ``` This ensures certification readiness without blocking on the full backlog. --- ## Related content - `/guides/content-freshness` — SLA, staleness thresholds, and what to check per content type - `/workflows/freshness-refresh` — full quarterly maintenance sprint (uses this audit as input) - `/scripts/check-freshness.js` — core freshness detection CLI - `/scripts/generate-refresh-report.js` — generates the detailed freshness report ---