Skill Issues: Compromising Claude Code with malicious skills & agents -- Part 1

This technical blog post demonstrates how attackers can compromise Claude Code, Anthropic's AI coding assistant, through malicious skill files and sub-agents. Skills are markdown files that instruct LLMs on how to perform specific tasks, and thousands of users share them on GitHub and skills.sh without proper vetting. The author shows that with default settings, a skill containing frontmatter with "allowed-tools: Bash(*)" and a dynamic context command (using !`command`) can execute arbitrary bash commands, including a reverse shell, without any user prompt or LLM reasoning. Sub-agents, which can run with "bypassPermissions" mode, can also execute malicious commands, such as installing a backdoored npm package. The article notes that while Claude Code has complex permission and command-parsing logic, the LLM itself may reject obviously malicious commands, but dynamic context inputs bypass this reasoning entirely. Defensive measures include denying Bash commands in settings files, sandboxing, carefully reviewing all skills and agents, and maintaining internal trusted sources. The author concludes that running untrusted skills carries similar risk to installing arbitrary pip packages or executables, and organizations must implement strong controls around AI coding tools. 

https://labs.reversec.com/posts/2026/05/skill-issues-compromising-claude-code-with-malicious-skills-agents-part-1

Comments

Post a Comment

Popular posts from this blog

Prompt Engineering Demands Rigorous Evaluation

In the blog post “Prompt Engineering Requires Evaluation” on the Shostack + Associates website, the author argues that treating prompts for large language models (LLMs) merely as creative artefacts is insufficient. Engineering prompts properly demands structured evaluation frameworks — what the AI community calls “evals” — to test which prompt versions work better, with which models, and under which conditions. The post highlights that simply assuming a prompt is “good enough” creates risks when LLMs are integrated into production systems (e.g., for threat modeling). It advocates for measuring prompt performance, variation effects, and tool-chain dependencies (model, context, ancillary materials).  Ultimately the message is: prompt engineering should borrow disciplined practices from software engineering (versioning, testing, benchmarking) rather than relying on informal experimentation.  https://shostack.org/blog/prompt-enignieering-requires-evaluation/
Read more

SecObserve: Simplified Vulnerability and License Management for CI/CD Pipelines

 SecObserve is an open-source tool for managing vulnerabilities and licenses in software development and cloud environments. It integrates various vulnerability scanners into CI/CD pipelines using GitLab CI templates and GitHub Actions for streamlined setup. It offers a centralized dashboard for assessing and reporting vulnerabilities, with tools for filtering, sorting, and evaluating results. SecObserve supports automation and manual assessments to focus on resolving critical issues.  https://github.com/MaibornWolff/SecObserve/tree/dev
Read more

NIST Updates Guidelines: Focus on Strong Passwords and MFA Over Frequent Rotation

 NIST has updated its guidelines, advising against mandatory password changes every 30-90 days unless a breach occurs. Frequent changes often lead to weak passwords, as users may make minimal adjustments. The focus has shifted to strong passwords and Multi-Factor Authentication (MFA) as more effective security measures. Despite this, automated password rotation remains crucial for securing sensitive accounts, especially for privileged users. It helps prevent unauthorized access, reduces exposure time, and ensures strong, unique passwords without burdening users. https://www.techradar.com/pro/navigating-nists-updated-password-rotation-guidelines
Read more