Agent skill

implementing-epss-score-for-vulnerability-prioritization

Integrate FIRST's Exploit Prediction Scoring System (EPSS) API to prioritize vulnerability remediation based on real-world exploitation probability within 30 days.

View SKILL.md on GitHub Repository
Stars 4,300
Forks 470

Install this agent skill to your Project

npx add-skill https://github.com/mukul975/Anthropic-Cybersecurity-Skills/tree/main/skills/implementing-epss-score-for-vulnerability-prioritization

SKILL.md

Implementing EPSS Score for Vulnerability Prioritization

Overview

The Exploit Prediction Scoring System (EPSS) is a data-driven model developed by FIRST (Forum of Incident Response and Security Teams) that estimates the probability of a CVE being exploited in the wild within the next 30 days. EPSS produces scores from 0.0 to 1.0 (0% to 100%) using machine learning trained on real-world exploitation data. Unlike CVSS which measures severity, EPSS measures likelihood of exploitation, making it essential for risk-based vulnerability prioritization.

When to Use

  • When deploying or configuring implementing epss score for vulnerability prioritization capabilities in your environment
  • When establishing security controls aligned to compliance requirements
  • When building or improving security architecture for this domain
  • When conducting security assessments that require this implementation

Prerequisites

  • Python 3.9+ with requests, pandas, matplotlib
  • Access to FIRST EPSS API (https://api.first.org/data/v1/epss)
  • Vulnerability scan results with CVE identifiers
  • Optional: NVD API key for CVSS enrichment

EPSS API Usage

Query Single CVE

bash
# Get EPSS score for a specific CVE
curl -s "https://api.first.org/data/v1/epss?cve=CVE-2024-3400" | python3 -m json.tool

# Response:
# {
#   "status": "OK",
#   "status-code": 200,
#   "version": "1.0",
#   "total": 1,
#   "data": [
#     {
#       "cve": "CVE-2024-3400",
#       "epss": "0.95732",
#       "percentile": "0.99721",
#       "date": "2024-04-15"
#     }
#   ]
# }

Query Multiple CVEs

bash
# Batch query up to 100 CVEs
curl -s "https://api.first.org/data/v1/epss?cve=CVE-2024-3400,CVE-2024-21887,CVE-2023-44228" | \
  python3 -c "
import sys, json
data = json.load(sys.stdin)
for item in data['data']:
    pct = float(item['epss']) * 100
    print(f\"{item['cve']}: {pct:.2f}% exploitation probability (percentile: {item['percentile']})\")
"

Download Full EPSS Dataset

bash
# Download complete daily EPSS scores (CSV format)
curl -s "https://epss.cyentia.com/epss_scores-current.csv.gz" | gunzip > epss_scores_current.csv

# Check size and preview
wc -l epss_scores_current.csv
head -5 epss_scores_current.csv

Query Historical EPSS Scores

bash
# Get EPSS score for a specific date
curl -s "https://api.first.org/data/v1/epss?cve=CVE-2024-3400&date=2024-04-12"

# Get time series data
curl -s "https://api.first.org/data/v1/epss?cve=CVE-2024-3400&scope=time-series"

Prioritization Strategy

EPSS + CVSS Combined Approach

EPSS Score CVSS Score Priority Action
> 0.7 >= 9.0 P0 - Immediate Remediate within 24 hours
> 0.7 >= 7.0 P1 - Urgent Remediate within 48 hours
> 0.4 >= 7.0 P2 - High Remediate within 7 days
> 0.1 >= 4.0 P3 - Medium Remediate within 30 days
<= 0.1 >= 7.0 P3 - Medium Remediate within 30 days
<= 0.1 < 7.0 P4 - Low Remediate within 90 days

EPSS Percentile Thresholds

  • Top 1% (percentile >= 0.99): Extremely likely to be exploited; treat as Critical
  • Top 5% (percentile >= 0.95): High exploitation probability; prioritize remediation
  • Top 10% (percentile >= 0.90): Elevated risk; schedule for near-term remediation
  • Bottom 50%: Low exploitation probability; handle in normal patch cycle

Implementation

python
import requests
import pandas as pd
from datetime import datetime

def fetch_epss_scores(cve_list):
    """Fetch EPSS scores for a list of CVEs from FIRST API."""
    scores = {}
    batch_size = 100
    for i in range(0, len(cve_list), batch_size):
        batch = cve_list[i:i + batch_size]
        resp = requests.get(
            "https://api.first.org/data/v1/epss",
            params={"cve": ",".join(batch)},
            timeout=30
        )
        if resp.status_code == 200:
            for entry in resp.json().get("data", []):
                scores[entry["cve"]] = {
                    "epss": float(entry["epss"]),
                    "percentile": float(entry["percentile"]),
                    "date": entry.get("date", ""),
                }
    return scores

def prioritize_vulnerabilities(scan_results_csv, output_csv):
    """Enrich scan results with EPSS scores and assign priorities."""
    df = pd.read_csv(scan_results_csv)
    cve_list = df["cve_id"].dropna().unique().tolist()

    epss_data = fetch_epss_scores(cve_list)

    df["epss_score"] = df["cve_id"].map(lambda c: epss_data.get(c, {}).get("epss", 0))
    df["epss_percentile"] = df["cve_id"].map(lambda c: epss_data.get(c, {}).get("percentile", 0))

    def assign_priority(row):
        epss = row.get("epss_score", 0)
        cvss = row.get("cvss_score", 0)
        if epss > 0.7 and cvss >= 9.0:
            return "P0"
        if epss > 0.7 and cvss >= 7.0:
            return "P1"
        if epss > 0.4 and cvss >= 7.0:
            return "P2"
        if epss > 0.1 or cvss >= 7.0:
            return "P3"
        return "P4"

    df["priority"] = df.apply(assign_priority, axis=1)
    df = df.sort_values(["priority", "epss_score"], ascending=[True, False])
    df.to_csv(output_csv, index=False)
    print(f"[+] Prioritized {len(df)} vulnerabilities -> {output_csv}")
    print(f"    P0: {len(df[df['priority']=='P0'])}")
    print(f"    P1: {len(df[df['priority']=='P1'])}")
    print(f"    P2: {len(df[df['priority']=='P2'])}")
    print(f"    P3: {len(df[df['priority']=='P3'])}")
    print(f"    P4: {len(df[df['priority']=='P4'])}")
    return df

EPSS Trend Analysis

python
def fetch_epss_timeseries(cve_id):
    """Get historical EPSS scores for trend analysis."""
    resp = requests.get(
        "https://api.first.org/data/v1/epss",
        params={"cve": cve_id, "scope": "time-series"},
        timeout=30
    )
    if resp.status_code == 200:
        return resp.json().get("data", [])
    return []

def detect_epss_spikes(cve_id, threshold=0.3):
    """Detect significant EPSS score increases indicating emerging threats."""
    timeseries = fetch_epss_timeseries(cve_id)
    if len(timeseries) < 2:
        return False
    sorted_data = sorted(timeseries, key=lambda x: x.get("date", ""))
    latest = float(sorted_data[-1].get("epss", 0))
    previous = float(sorted_data[-2].get("epss", 0))
    increase = latest - previous
    if increase >= threshold:
        print(f"[!] EPSS spike detected for {cve_id}: {previous:.3f} -> {latest:.3f} (+{increase:.3f})")
        return True
    return False

References

Recommended Agent Skills

Expand your agent's capabilities with these related and highly-rated skills.

mukul975/Anthropic-Cybersecurity-Skills

mapping-mitre-attack-techniques

Maps observed adversary behaviors, security alerts, and detection rules to MITRE ATT&CK techniques and sub-techniques to quantify detection coverage and guide control prioritization. Use when building an ATT&CK-based coverage heatmap, tagging SIEM alerts with technique IDs, aligning security controls to adversary playbooks, or reporting threat exposure to executives. Activates for requests involving ATT&CK Navigator, Sigma rules, MITRE D3FEND, or coverage gap analysis.

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

hunting-for-spearphishing-indicators

Hunt for spearphishing campaign indicators across email logs, endpoint telemetry, and network data to detect targeted email attacks.

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

analyzing-malicious-url-with-urlscan

URLScan.io is a free service for scanning and analyzing suspicious URLs. It captures screenshots, DOM content, HTTP transactions, JavaScript behavior, and network connections of web pages in an isolat

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

implementing-zero-standing-privilege-with-cyberark

Deploy CyberArk Secure Cloud Access to eliminate standing privileges in hybrid and multi-cloud environments using just-in-time access with time, entitlement, and approval controls.

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

implementing-pam-for-database-access

Deploy privileged access management for database systems including Oracle, SQL Server, PostgreSQL, and MySQL. Covers session proxy configuration, credential vaulting, query auditing, dynamic credentia

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

detecting-t1003-credential-dumping-with-edr

Detect OS credential dumping techniques targeting LSASS memory, SAM database, NTDS.dit, and cached credentials using EDR telemetry, Sysmon process access monitoring, and Windows security event correlation.

4,300 470
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results