The Agent Skills Directory

Execution of User-Provided Code: The skill involves cloning and running code from external GitHub repositories provided in bug reports. While this introduces a potential security consideration, the skill includes explicit instructions for the agent to inspect the source code for suspicious patterns, such as postinstall scripts or obfuscated code, before proceeding.
Package Installation and Execution: The skill performs standard development operations, including npm install for Playwright packages and running tests via npm run ctest. These actions are conducted within a dedicated temporary directory (~/tmp/issue-<number>/) and are consistent with standard bug reproduction workflows.
External Data Processing: The agent processes content from GitHub issues, which represents an indirect prompt injection surface. The skill mitigates this by providing a structured triage workflow and specific safety guidance for evaluating untrusted input.

playwright-triage