Runtime Skill Audit: Targeted Runtime Probing for Agent Skill Security

Agent skills let LLM agents reuse instructions, resources, tools, and workflows, but they also create a new place for malicious behavior to hide. A skill may look benign in its documentation or code while becoming harmful only when it is invoked with particular user requests, local assets,...

offensive-claude-533

Offensive Security Research Config for Claude Code !TIP...

offensive-claude-604

Offensive Security Research Config for Claude Code !TIP...

offensive-claude-813

Offensive Security Research Config for Claude Code !TIP...

offensive-claude-982

Offensive Security Research Config for Claude Code !TIP...

offensive-claude-490

Offensive Security Research Config for Claude Code !TIP...

CVE-2026-10213

A security flaw has been discovered in AstrBotDevs AstrBot 4.23.6. This vulnerability affects unknown code of the file /api/skills/delete of the component API Endpoint. Performing a manipulation of the argument Name results in path traversal. The attack can be initiated remotely. The exploit has...

CVE-2026-7784

A vulnerability has been found in RTGS2017 NagaAgent up to 5.1.0. This issue affects some unknown processing of the file apiserver/routes/extensions.py of the component Skills Endpoint. Such manipulation of the argument Name leads to path traversal. It is possible to launch the attack remotely. T...

CVE-2026-41863

Spring AI's support for Anthropic's Skills API used LLM-influenced filenames unsanitized in Path.resolve before writing files to disk. This could allow a malicious user to write files outside the intended target directory, including restricted directories. Affected versions: Spring AI: 1.1.0...

MalSkillBench: A Runtime-Verified Benchmark of Malicious Agent Skills

AI coding agents such as Claude Code and Gemini CLI increasingly extend themselves with third-party skills: markdown packages bundling natural-language instructions, executable scripts, and tool permissions. Because a skill is at once code and agent-facing instruction, it introduces a supply chai...

Winning the cyber marathon with Tony Giandomenico

In the high-speed world of cybersecurity, the difference between a breach and a breakthrough often comes down to endurance. Tony Giandomenico, Senior Director of Product Management with Cisco Talos, joins me to discuss how he balances the intensity of leading major product launches with the...

CVE-2026-10220

A vulnerability was determined in NousResearch hermes-agent up to 2026.4.30. Affected is the function servepluginskill/skillview of the file tools/skillstool.py. Executing a manipulation can lead to injection. The attack may be performed from remote. The exploit has been publicly disclosed and ma...

CVE-2026-10220 NousResearch hermes-agent skills_tool.py skill_view injection

A vulnerability was determined in NousResearch hermes-agent up to 2026.4.30. Affected is the function servepluginskill/skillview of the file tools/skillstool.py. Executing a manipulation can lead to injection. The attack may be performed from remote. The exploit has been publicly disclosed and ma...

CVE-2026-10220 NousResearch hermes-agent skills_tool.py skill_view injection

A vulnerability was determined in NousResearch hermes-agent up to 2026.4.30. Affected is the function servepluginskill/skillview of the file tools/skillstool.py. Executing a manipulation can lead to injection. The attack may be performed from remote. The exploit has been publicly disclosed and ma...

CVE-2026-10220

CVE-2026-10220 affects NousResearch hermes-agent up to version 2026.4.30. The vulnerability targets the function _serve_plugin_skill/skill_view in tools/skills_tool.py, where a manipulation can cause injection. It is described as a remote-access issue with a publicly disclosed exploit (PoC). The ...

CVE-2026-10213

A security flaw has been discovered in AstrBotDevs AstrBot 4.23.6. This vulnerability affects unknown code of the file /api/skills/delete of the component API Endpoint. Performing a manipulation of the argument Name results in path traversal. The attack can be initiated remotely. The exploit has...

CVE-2026-10213

A security flaw has been discovered in AstrBotDevs AstrBot 4.23.6. This vulnerability affects unknown code of the file /api/skills/delete of the component API Endpoint. Performing a manipulation of the argument Name results in path traversal. The attack can be initiated remotely. The exploit has...

EUVD-2026-33534

A security flaw has been discovered in AstrBotDevs AstrBot 4.23.6. This vulnerability affects unknown code of the file /api/skills/delete of the component API Endpoint. Performing a manipulation of the argument Name results in path traversal. The attack can be initiated remotely. The exploit has...

AstrBot 路径遍历漏洞

AstrBot is an open-source multi-platform LLM chatbot and development framework created by AstrBot. Version 4.23.6 of AstrBot contains a path traversal vulnerability. This vulnerability stems from improper handling of the Name parameter in the/api/skills/delete file within the API Endpoint...

SkillGuard: A Permission Framework for Agent Skills

Agent skills extend LLM agents with reusable instructions, scripts, tool bindings, and contextual dependencies. However, current skill ecosystems largely rely on trust-based loading and static inspection, leaving a gap between what a skill can inject into an agent's context and what it can cause...