The Agent Skills Directory

[SAFE]: The skill instructions and associated scripts align with the stated purpose of bug diagnosis. No malicious patterns such as prompt injection, data exfiltration, or obfuscation were detected.
[COMMAND_EXECUTION]: The methodology involves executing local commands (e.g., tests, curl, CLI invocations) to create a feedback loop. These operations are standard for a debugging workflow and are intended to be run within the local development environment.
[EXTERNAL_DOWNLOADS]: The skill mentions using well-known tools like Playwright and Puppeteer for headless browser testing. These are industry-standard utilities referenced neutrally as part of the suggested debugging toolkit.

diagnosing-bugs