Injection Vulnerability in NousResearch Hermes Agent Plugin Handler

A vulnerability was determined in NousResearch hermes-agent up to 2026.4.30. Affected is the function _serve_plugin_skill/skill_view of the file tools/skills_tool.py. Executing a manipulation can lead to injection. The attack may be performed from remote. The exploit has been publicly disclosed and may be utilized. The vendor was contacted early about this disclosure but did not respond in any way.

Casky was already ahead

This CVE exploits attack patterns that Casky's 301matched skills already investigate — long before this vulnerability was disclosed. Claude's reasoning model maps these techniques to MITRE ATT&CK, so practitioners who ran these skills have already seen the threat behaviour in their findings.

Analysis

CVE-2026-10220 represents a critical injection vulnerability in NousResearch's hermes-agent framework, specifically within the plugin skill serving mechanism. The _serve_plugin_skill/skill_view function in tools/skills_tool.py fails to properly sanitize user-supplied input before processing, allowing attackers to inject malicious code or commands remotely. This affects organizations deploying hermes-agent versions up to 2026.4.30 for AI-powered agent automation and skill orchestration. The vulnerability is particularly concerning because it enables unauthenticated remote exploitation, and public disclosure combined with vendor non-responsiveness increases the attack surface for threat actors seeking to compromise AI agent infrastructure.

Casky.ai's 301 mapped skills detect this vulnerability through behavioral analysis aligned with MITRE ATT&CK techniques TA0001 (Reconnaissance) and TA0043 (Reconnaissance). Practitioners using Casky would observe detection across multiple skill categories: input validation bypass patterns (CWE-74), improper neutralization techniques (CWE-707), and malicious skill parameter injection attempts. Claude's extended reasoning capabilities enable detection of subtle injection variants—including polyglot payloads and encoding evasion—that traditional signature-based tools miss. Security teams would see findings highlighting suspicious plugin invocations, unusual parameter construction in skill_view calls, and reconnaissance activity probing the plugin handler for injection points, allowing them to identify exploitation attempts before payload execution.

Live Threat IntelligenceComing soon

Casky Risk Score

—

EPSS Probability

—

Shodan Exposure

—

GreyNoise Signal

—

Composite risk scoring from EPSS, CISA KEV, Shodan, and GreyNoise — 21 security APIs correlated into a single Casky Risk Score. Coming in Casky Pro. Join early access →

301 Casky skills cover this attack pattern

These skills use Claude AI's reasoning model to surface findings in the same attack categories as CVE-2026-10220.

analyzing-apt-group-with-mitre-navigator

threat intelligence · low

Run

analyzing-campaign-attribution-evidence

threat intelligence · low

Run

analyzing-certificate-transparency-for-phishing

threat intelligence · low

Run

MITRE ATT&CK Techniques

TA0001 TA0043

Investigate the attack patterns behind CVE-2026-10220

Casky has 301 skills that investigate the attack patterns behind CVE-2026-10220. Run one and get CVSS-scored findings in 3 minutes.

Run the skill that detects this →

X Instagram LinkedIn