Works with the AI Agents you already use
Vulnerability Patch Triage: KEV, EPSS and Patch SLAs
Too many CVEs, not enough time? Rank scanner findings by real risk (CISA KEV, FIRST EPSS, your exposure and asset criticality), set P0-P3 deadlines and plan no-fix controls. ACT NOW / SCHEDULE / MONITOR.
$14
Vulnerability Patch Triage: KEV, EPSS and Patch SLAs
Example session with this skill installed
Our Tenable export has 1,900 findings across 120 assets after the PHP2Shell week. I've attached it plus our asset list with internet-facing and criticality tags. What do we patch first?
- Read your context and instructions
- Compiled the vulnerability patch triage
Patch triage: production + corporate estate (2026-10-23)
Data sources: Tenable export (1,906 rows, expanded to 2,341 CVE-asset pairs), asset list (120 assets), CISA KEV (fetched today), FIRST EPSS (scores dated today).
Summary: P0 14, P1 63, P2 412, P3 1,852. SLA breaches 9. No-fix items 6.
Act now (P0, sample)
| Pri | CVE | Asset | Why | Fix or control | Owner | Due |
|---|---|---|---|---|---|---|
| P0 | CVE-2026-109618 | web-01..04 | KEV; internet-facing; customer data | PHP 8.3.36 + restart FPM and workers | Web | Oct 24 |
| P0 | CVE-2026-109619 | legacy-01 | KEV; internet-facing; no fix for 7.4 | No-fix plan: VPN-only access, Cloudflare WAF, uploads disabled; migrate to 8.3 | Web | Contain Oct 23 |
| P0 | (KEV) VPN appliance CVE | vpn-01 | KEV; internet-facing; identity | Vendor firmware update | Network | Oct 24 |
| P1 | (EPSS 0.41) CMS plugin CVE | marketing-wp | Public exploit; internet-facing | Plugin update | Marketing | Oct 30 |
Biggest wins (batched by component)
- php upgrade on 9 hosts closes 31 findings (all P0/P1)
- openssl package update on 44 Linux hosts closes 210 findings (P2)
- Chrome auto-update policy closes 380 findings (P3)
Overrides (with reasons)
- 3 log4j findings downgraded P1 -> P3: the jar is present but not loaded (confirmed by
lsofand app config). Re-check after the next deploy.
SLA breaches
9 items, all P2 on build servers older than 30 days. Owners notified, with a new due date of Nov 6.
Exceptions register
- legacy-01 PHP 7.4: accepted until Oct 31 with VPN-only access and WAF. Owner: Head of Web.
Changes since last run
- 2 CVEs added to KEV today: moved P2 -> P0.
Connects securely to your tools. The creator never sees your data.
About this skill
AI tools are now finding vulnerabilities faster than teams can patch them. October 2026's PHP2Shell flaws sat unnoticed for over a decade until AI-assisted research surfaced them, and AI-assisted research is now surfacing flaws across open source. Your scanner says you have 2,400 "critical" findings. Sorting by CVSS doesn't tell you which ones attackers are actually using, or which ones sit on an internet-facing server holding customer data.
This skill gives your AI agent a risk-based triage method and a triage script that combines the signals that matter. It:
- Takes scanner exports (Tenable, Qualys, Rapid7, Wiz, Snyk, Dependabot, Trivy, Grype), SBOM matches or a pasted CVE list, and normalises them into one table
- Enriches each CVE live with the CISA KEV catalog (known exploited vulnerabilities) and FIRST EPSS scores (probability of exploitation in the next 30 days). It never states these from memory
- Asks for the context only you have: is it internet-facing, and how critical is the asset?
- Applies an SSVC-style decision model to assign P0-P3 with default deadlines (3/7/30/90 days, editable) and a one-line reason for every urgent item, so humans can challenge it
- Supports overrides with written reasons, such as reachability analysis, components not deployed or features disabled
- Builds no-fix plans from a compensating-controls playbook: remove exposure, disable the feature, vendor mitigations (marked partial), WAF virtual patching (marked as only covering routed traffic), segmentation, detection and re-check dates
- Groups fixes by component and owner, since one package upgrade often closes dozens of findings
- Tracks SLA breaches and keeps an exceptions register with owners, compensating controls and expiry dates
- Highlights items whose KEV or EPSS status changed since the last run
For security engineers, IT and DevOps teams, MSPs and small companies without a full vulnerability-management team.
Defensive and honest: no exploit code, no invented scores, unknowns clearly marked, and a reminder that scores support judgement rather than guarantee a low-priority item won't be exploited. Pairs with the PHP2Shell Patch and Exposure Check skill.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.
Prioritisation depends on accurate exposure and criticality data from you. KEV/EPSS enrichment needs network access. Scores support judgement and do not guarantee safety. Defensive only: no exploit code or exploitability testing.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 24 days ago
- Accounts
Cisa
First
- Software
Python 3
- Optional
- Internet access
Python runs the triage script. You provide a scanner CSV export and an asset list with exposure and criticality. Live enrichment calls cisa.gov and api.first.org.
Permissions
Allowed Hosts