pa-knowledge-base-curator — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited pa-knowledge-base-curator (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Maintains alignment between product capabilities and customer-facing documentation. Identifies content gaps from support patterns, detects stale content after product changes, and surfaces relevant resources to customers proactively. Part of the Product Adoption & Value Realisation pillar.
This is a monitoring and recommendation skill. It identifies where documentation is missing, outdated, or underperforming. It does not write the documentation (that is content team or enablement work) -- it tells you what needs writing, what needs updating, and estimates the impact of fixing it.
Analyse support ticket data for signals that documentation is missing or inadequate:
| Signal | Detection Criteria | Confidence | Estimated Gap Impact |
|---|---|---|---|
| Recurring question cluster | 5+ tickets from different accounts on the same topic in 30 days | High | High -- same question being asked repeatedly means the answer is not findable |
| "How do I" language | Tickets phrased as procedural questions ("How do I set up...", "Where can I find...") | High | High -- these are documentation-solvable questions |
| High-traffic article followed by ticket | Article viewed and then a ticket created within 30 minutes by the same user | High | High -- the article exists but does not answer the question |
| Search-with-no-result | Help centre search queries that return zero matches (if analytics available) | Medium | Medium -- demand exists for content that does not |
| New feature tickets | Spike in tickets about a feature within 30 days of its release, with no corresponding help article | High | High -- documentation was not created alongside the release |
| Workaround descriptions | Tickets where the CSM or support agent describes a workaround for something that should be documented | Medium | Medium -- the knowledge exists internally but is not published |
For each gap identified, compute:
After each product release:
| Impact | Criteria | Priority |
|---|---|---|
| Breaking | Article references UI, workflow, or feature that has been removed or fundamentally changed | Critical -- article will actively mislead readers |
| Outdated | Article is factually correct but references old UI or does not cover new capabilities | High -- article is incomplete |
| Minor | Article is mostly correct but could mention a new option or improvement | Low -- update when convenient |
| Unaffected | Article covers content that was not changed in this release | None -- no action needed |
Based on a customer's adoption profile from pa-adoption-tracker:
Track whether content is actually solving problems:
| Metric | Computation | What It Tells You |
|---|---|---|
| View-to-ticket ratio | Tickets created within 30 min of article view / total article views | Whether the article answers the question (lower is better) |
| Search satisfaction | Searches that result in an article click (not a "no results" or immediate bounce) | Whether content is findable |
| Content gap ticket volume | Total tickets in the last 30 days that match identified gaps | The cost of not having the content |
| Post-publication deflection | Ticket volume on a topic before vs. after documentation was published | Whether the new content actually reduced tickets |
| Staleness rate | Percentage of articles not reviewed since the last 2 product releases | How much of the knowledge base is at risk of being outdated |
Weekly content gap report:
{
"report_date": "2026-03-10",
"gaps_identified": [
{
"topic": "Custom API authentication for SSO environments",
"evidence": "12 tickets from 8 accounts in 30 days. All procedural 'how do I' questions",
"current_article": "none",
"estimated_ticket_deflection_monthly": 10,
"arr_of_affected_accounts": 420000,
"recommended_content_type": "Step-by-step guide with screenshots",
"estimated_creation_effort": "2-3 hours",
"priority": "high"
},
{
"topic": "Bulk import performance with large datasets",
"evidence": "6 tickets from 5 accounts. Current article covers basic import but not performance optimisation for large files",
"current_article": "KB-2025-0089",
"article_monthly_views": 180,
"view_to_ticket_ratio": 0.033,
"recommended_action": "Expand existing article with performance section",
"estimated_effort": "1 hour",
"priority": "medium"
}
],
"stale_articles": [
{
"article_id": "KB-2024-0142",
"title": "Setting up dashboard widgets",
"last_updated": "2025-06-15",
"product_changes_since": 2,
"impact": "outdated",
"monthly_views": 340,
"priority": "high"
}
],
"content_health_summary": {
"total_articles": 142,
"current": 98,
"needs_update": 31,
"outdated_critical": 4,
"staleness_rate": 0.25,
"avg_view_to_ticket_ratio": 0.018,
"estimated_monthly_deflectable_tickets": 28
}
}Per-customer resource recommendations (for CSM or pa-enablement-orchestrator):
{
"account_id": "string",
"recommendations": [
{
"feature": "Advanced Reporting",
"adoption_status": "untouched",
"segment_adoption_rate": 0.78,
"resources": [
{ "title": "Getting Started with Custom Reports", "type": "tutorial", "url": "https://...", "duration": "5 min read" },
{ "title": "Custom Report Templates Gallery", "type": "reference", "url": "https://...", "duration": "browse" }
],
"suggested_framing": "Your team has access to custom reporting -- this 5-minute guide covers the basics. Most teams at your tier use this to cut manual reporting time"
}
]
}| Decision Point | Context Provided | What the Human Decides |
|---|---|---|
| Gap prioritisation | Ranked gaps with ticket volume, ARR weight, and effort estimates | Which gaps to address first, given content team capacity. The skill ranks by impact; the content team may have capacity constraints |
| Content creation | Gap specification with evidence and content type recommendation | The actual content. Writing help articles is human work -- the skill defines what is needed, not how to write it |
| Customer resource delivery | Per-customer recommendations with framing suggestions | Whether and how to share resources with specific customers. The CSM knows whether sending a help article will be helpful or patronising |
| Stale content triage | Stale articles ranked by traffic and severity | Whether to update, archive, or redirect each stale article. Some outdated articles are still useful with a minor edit; others should be retired |
Required:
Strongly recommended:
Downstream consumers:
references/content-audit-process.md -- Post-release audit process, quarterly full audit, and gap report methodology~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.