The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/capture.ts dynamically loads and executes local JavaScript or TypeScript files using the import() function based on the --pre-script command-line argument. This enables the execution of arbitrary code from file paths specified at runtime.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes and extracts text from external web pages. Ingestion points: Untrusted data enters the agent context via page.goto() and subsequent DOM text extraction in scripts/capture.ts. Boundary markers: There are no specific delimiters or warnings used to wrap the extracted text to prevent instructions from being followed. Capability inventory: The skill has the ability to write to the file system, execute local scripts dynamically, and access the network via a browser. Sanitization: The skill implements regex-based redaction in scripts/capture.ts and a dedicated validation script scripts/validate-sensitive.ts to identify and mask common secret patterns before data is stored or uploaded.

screengrabs