The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses various shell commands (find, grep, xargs, sed) to identify project structure and perform bulk updates to import statements in test files. It also uses package manager commands (npm, pnpm, yarn) to install dependencies.
[EXTERNAL_DOWNLOADS]: The skill automates the installation of Playwright and Stably SDK packages from official registries. These include @playwright/test, @stablyai/playwright-test, @stablyai/email, and @playwright/mcp. These resources are provided by the skill vendor or well-known organizations.
[DYNAMIC_EXECUTION]: The skill generates a local verification test file (stably-verification.spec.ts) and executes it using the project's Playwright runner to confirm successful installation.
[PRIVILEGE_ESCALATION]: The skill includes an optional step to perform a global installation of the Playwright MCP tool (npm install -g @playwright/mcp). The agent is instructed to ask for explicit user permission before proceeding with this action.
[INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface as it reads project metadata from package.json, configuration from playwright.config.ts, and identifies test files. It relies on standard delimiters and does not process untrusted external data other than the project's own source code.
[SAFE_PRACTICE]: The skill contains explicit instructions to protect sensitive data, stating that the agent should never read or write .env files directly and must instead guide the user to add credentials manually.

stably-sdk-setup