The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions for the user to execute an external script directly in their shell using the pattern curl -fsSL https://www.lazyweb.com/install.sh | bash. This is a high-risk security practice that allows arbitrary code execution from a remote source without verification.\n- [COMMAND_EXECUTION]: The skill instructs the agent to run local shell commands, such as using cat to read version information from the ~/.lazyweb/ directory and suggesting the execution of an unverified installation script.\n- [EXTERNAL_DOWNLOADS]: The core logic of the skill involves fetching and processing instruction files from over 50 different external GitHub repositories. While some of these belong to trusted organizations, many are maintained by individual community members or unverified third-party accounts, introducing a significant supply chain risk.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its architecture of fetching and following instructions from external sources.\n
Ingestion points: The skill retrieves SKILL.md files from various third-party URLs (e.g., github.com/emilkowalski/skill, github.com/Dammyjay93/interface-design).\n
Boundary markers: The instructions lack technical delimiters or explicit warnings to the agent to treat external content as untrusted data.\n
Capability inventory: The skill allows the use of powerful tools including Bash, Write, and Agent, which could be exploited if malicious instructions are injected via the fetched files.\n
Sanitization: There is no automated verification, validation, or filtering of the content retrieved from these external sources before the agent incorporates it into its workflow.

lazyweb-design-best-practices