The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and analyze HTML and JavaScript files from user-provided target URLs. This untrusted content could potentially contain malicious instructions intended to influence the agent's behavior during the extraction process.
Ingestion points: Multiple curl commands in SKILL.md and the js_extract.sh script in references/js-extract-patterns.md download content from external domains.
Boundary markers: Absent. There are no explicit delimiters or instructions to the agent to treat the downloaded JS/HTML content as passive data.
Capability inventory: The skill uses shell commands (curl, grep, sed, sort) to perform network requests and data processing.
Sanitization: Absent. The downloaded content is processed as raw text using regular expressions without additional sanitization layers.
[COMMAND_EXECUTION]: The skill instructs the agent to execute a variety of shell commands and provides a complete bash script for automation. These commands involve network operations and local file system management (mkdir, cat) to store and process reconnaissance results.

js-api-extract