The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to browse and interpret third-party websites, which creates a significant surface for indirect prompt injection. Malicious content on a target website could attempt to hijack the agent's instructions, manipulate test outcomes, or compromise the integrity of the generated reports. Ingestion points: Page content, DOM snapshots, console logs, and network responses are ingested into the agent's context in several modes. Boundary markers: The documentation does not specify the use of delimiters or instructions to ignore embedded commands within the site data. Capability inventory: The agent can write to the local filesystem and execute shell commands. Sanitization: No sanitization of web content is specified before processing.
[EXTERNAL_DOWNLOADS]: The scripts/superqa.sh script is configured to automatically download and install Python packages (textual, playwright, pyyaml, pillow) from the public PyPI registry and the Chromium browser binary from Playwright's infrastructure at runtime.
[COMMAND_EXECUTION]: The skill uses shell scripts and CLI tools for core functionality, including launching a terminal TUI, running test scenarios via Python modules, and managing browser instances. It executes both shell scripts and Windows PowerShell scripts to facilitate these operations.

browser-qa