The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses standard package manager commands (npm init playwright, npm install, npx playwright install) to set up the testing environment. These are necessary for the skill's primary purpose.
[EXTERNAL_DOWNLOADS]: Downloads the official Playwright framework and browser binaries from well-known repositories. It also generates GitHub Actions workflows that reference official, trusted actions from the actions organization.
[INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface as it scans project files like package.json and tsconfig.json to detect the environment.
Ingestion points: Reads package.json, tsconfig.json, and CI configuration files from the local project root.
Boundary markers: The global protocols section instructs the agent to treat external data as hostile and redact secrets, providing basic defensive context.
Capability inventory: The skill has file-writing capabilities and command execution through npm and npx.
Sanitization: No explicit sanitization of the values read from package.json is described, but the usage is limited to selecting configuration templates.

pw-init