Skill

performing-endpoint-vulnerability-remediation

skill:mukul975/anthropic-cybersecurity-skills#performing-endpoint-vulnerability-remediation

Performs vulnerability remediation on endpoints by prioritizing CVEs based on risk scoring, deploying patches, applying configuration changes, and validating fixes. Use when remediating findings from vulnerability scans, responding to critical CVE advisories, or maintaining endpoint compliance with patch management SLAs. Activates for requests involving vulnerability remediation, CVE patching, endpoint vulnerability management, or security fix deployment.

SkillsMP

Verdict

Suspicious

Risky behavior the analysis could not rule out.

Analyzed Oct 27, 2026, 23:34 UTC at 54a7988

findings

medium

highest severity

threat classes

4,532

files scanned

151/154

checks satisfied

package advisories

Why this verdict

suspicious: finding skill:skills/performing-endpoint-vulnerability-remediation:skills/performing-endpoint-vulnerability-remediation/SKILL.md:118:SUPPLY

Security findings 1

Supply chain

mediumContextual risklow confidence

The skill's patch-deployment step instructs installing the third-party PSWindowsUpdate module from the PowerShell Gallery with -Force and no pinned version, fetching unpinned remote code at execution time; it is disclosed and on-task for the declared patching purpose, so it is a contextual supply-chain risk rather than concealed behavior.

skills/performing-endpoint-vulnerability-remediation/SKILL.md:118-119

Install-Module PSWindowsUpdate -Force

vector:instruction

Counter-argument considered

Attempted refutation fails: the carve-out for installs requires pinned official-registry installs, but the skill body's command carries no version pin and uses -Force (skipping the gallery-trust prompt), and per the rules open documentation does not neutralize supply-chain harm; the command is nonetheless openly shown as a commented step in the declared patch-deployment workflow, so the finding's own contextual_risk characterization is accurate and it stands.

Open questions 2

What the analysis could not settle from the code alone.

The frontmatter maps this remediation skill to offensive ATT&CK techniques (T1055 process injection, T1547 boot autostart persistence, T1059 command shell, T1036 masquerading) that do not match the declared remediation purpose; is this inert documentation tagging or does any unit act on these mappings?

The claim 'No external packages — Python standard library only.' cited by the PY-CLAIM-CONTRADICTION detector leads does not appear in any provided unit (likely in the unshown SKILL.md lines 34-55); the lead cannot be confirmed, and note that socket/urllib use would not actually contradict a stdlib-only claim, so the flagged units (socket connect, CISA/EPSS fetches) were verified as declared-capability behavior rather than exfiltration.

Coverage gaps 4

Parts of the artifact the static analysis could not fully examine.

ledger_finding_without_valid_finding run 1: behavior-chains.0 SRC:ecb4c46f17d67f9c

ledger_finding_without_valid_finding run 1: behavior-chains.0 SRC:0a79902550e75df7

ledger_finding_without_valid_finding run 1: behavior-chains.0 SRC:0af14b4a65571efc

invalid_judge_finding behavior-chains.0