Skip to content

JPeetz/agent-skills / skills/security/vulnerability-management

vulnerability-management

Verified

Use this skill when implementing or operating a vulnerability management program — automated vulnerability scanning, risk-based prioritization (CVSS/EPSS), patch management automation, zero-day response procedures, CVE tracking and monitoring, vulnerability metrics tracking, remediation verification, security advisory monitoring, and threat intelligence integration.

FileKindSize
CHANGELOG.mdmarkdown1.8 KB
LICENSEother1.0 KB
SKILL.mdskill md21.6 KB
evals/evals.jsonconfig3.6 KB
references/sources.mdmarkdown7.3 KB
scripts/cve_lookup.pyexecutablescript7.7 KB
scripts/prioritize.pyexecutablescript8.7 KB

evals/evals.json

View on GitHub
{
  "skill": "vulnerability-management",
  "version": "1.0.0",
  "cases": [
    {
      "id": "eval-001-prioritize-noisy-backlog",
      "description": "Risk-based prioritization of a mixed scanner backlog where CVSS order and real urgency disagree.",
      "input": "Our scanner reports 412 findings: a CVSS 9.8 in an internal-only dev library with no public exploit, a CVSS 6.5 on an internet-facing VPN appliance listed in CISA KEV with EPSS 0.72, and a CVSS 7.5 on an internal Tier 2 app with no known exploit. What do we fix first and how should we prioritize the rest?",
      "shouldTrigger": true,
      "expect": "Skill activates and ranks by composite risk, not raw CVSS: KEV/EPSS internet-facing VPN finding first (emergency), internal 7.5 second, unreachable 9.8 last. Explains CVSS as intrinsic severity, EPSS as exploitation probability, KEV as confirmed exploitation, and applies exposure and asset criticality weights.",
      "assertions": [
        "Ranks the internet-facing KEV/EPSS finding above the CVSS 9.8 internal finding",
        "Explicitly distinguishes CVSS base severity from EPSS probability and KEV status",
        "Applies asset exposure and criticality to the ranking",
        "Gives concrete SLA deadlines per severity/exposure band"
      ]
    },
    {
      "id": "eval-002-zero-day-response",
      "description": "Zero-day with no vendor patch requires a containment-first playbook, not patching.",
      "input": "A critical unauthenticated RCE in our edge appliance was disclosed 3 hours ago. The vendor has no patch yet and there are reports of exploitation in the wild. We run 14 of these appliances, 6 internet-facing. What do we do right now?",
      "shouldTrigger": true,
      "expect": "Skill activates and walks the zero-day playbook: confirm the report, assign an incident commander, determine exposure, contain with feature disable/network block/WAF virtual patch and privilege reduction before any patching, hunt for prior compromise, increase logging, then patch on vendor release via emergency change with verification and credential rotation.",
      "assertions": [
        "Recommends containment before patching since no patch exists",
        "Names specific containment options (disable feature, segmentation, WAF rule)",
        "Includes compromise hunting and log/detection increases",
        "Mentions emergency change process and post-patch verification",
        "Includes a communication cadence and single authoritative channel"
      ]
    },
    {
      "id": "eval-003-patch-verification-and-metrics",
      "description": "Team claims remediation complete; agent must require verification evidence and define program metrics.",
      "input": "We deployed the patches for all 30 critical findings last night and closed the tickets. Can you help me report to leadership that we're at zero criticals, and set up metrics for next quarter?",
      "shouldTrigger": true,
      "expect": "Skill activates and refuses to certify zero criticals without re-scan verification (running version, not just installed package; container/JVM stale artifact risk). Defines metrics: MTTR, remediation rate, exposure window, SLA compliance, backlog aging, scan coverage, reopen rate, KEV exposure, and a monthly/quarterly reporting cadence.",
      "assertions": [
        "Requires re-scan verification before claiming closure",
        "Notes stale-artifact risk (containers, JVM apps, running vs installed version)",
        "Defines MTTR, SLA compliance, coverage, and reopen rate",
        "Warns against activity metrics and gamed MTTR",
        "Proposes a reporting cadence and audit evidence retention"
      ]
    }
  ]
}
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall