The Agent Skills Directory

[SAFE]: The skill's primary function is to perform security and usability audits on web applications. The methods described, including fault injection and network mocking, are standard industry practices for resilience testing.
[COMMAND_EXECUTION]: The skill utilizes browser-based automation tools (such as page.route for network interception and CODE blocks for scripting interactions). These operations are restricted to the context of the browser session and are necessary for the skill's stated purpose of simulating application failures.
[DATA_EXPOSURE]: There is a check (ERR-10) designed to detect if personally identifiable information (PII) is exposed in browser console logs. This is a defensive security audit check and does not involve exfiltrating data to an external source.
[INDIRECT_PROMPT_INJECTION]: As the skill interacts with external web pages and their content, it has an inherent surface for indirect prompt injection. However, it does not possess capabilities that would allow external content to trigger dangerous operations, and its use of structured reporting mitigates risk.

resilience-review