The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to perform extraction and metadata gathering. It constructs these commands by placing user-provided URLs directly into shell command strings (e.g., reader "$ARTICLE_URL"). This pattern is vulnerable to command injection if the URL contains shell metacharacters designed to break out of the quoted string and execute arbitrary commands.
[EXTERNAL_DOWNLOADS]: The instructions include steps to install external software packages from public registries if they are not present on the host system. Specifically, it suggests installing @mozilla/readability-cli, reader-cli, and trafilatura.
[REMOTE_CODE_EXECUTION]: The skill relies on the installation and execution of third-party packages from external registries. Automated installation of external code based on user-driven tasks introduces risks associated with package integrity and supply chain security.
[PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external web URLs, creating a surface for indirect prompt injection. * Ingestion points: Content is retrieved from external websites via curl, reader, and trafilatura (SKILL.md). * Capability inventory: The skill utilizes the Bash tool for command execution and the Write tool for file creation. * Sanitization: No sanitization or validation is performed on the URL or the retrieved HTML content before it is passed to shell commands or processing scripts. * Boundary markers: There are no explicit delimiters or instructions to treat the external content as untrusted data.

article-extractor