The Agent Skills Directory

[COMMAND_EXECUTION]: The provided bash scripts interpolate the user-supplied $ARTICLE_URL variable directly into shell commands (e.g., reader "$ARTICLE_URL"). This pattern creates a risk of command injection if the agent does not sanitize the input URL for shell metacharacters before execution.- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and display untrusted data from external websites, creating a surface for indirect prompt injection.
Ingestion points: External content is fetched from arbitrary URLs in SKILL.md using curl, reader, or trafilatura.
Boundary markers: No delimiters or safety instructions are used to distinguish external content from agent instructions.
Capability inventory: The skill utilizes Bash and Write tools to execute commands and modify the file system.
Sanitization: The extracted article text is saved and previewed without any sanitization or filtering of potentially malicious instructions embedded in the source webpage.- [EXTERNAL_DOWNLOADS]: The skill references and facilitates the installation of the @mozilla/readability-cli Node.js package and the trafilatura Python package from official registries.

article-extractor