The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run shell commands including git, grep, and npx. These commands are dynamically constructed using session-based variables and local environment paths.
[EXTERNAL_DOWNLOADS]: The skill utilizes npx --yes @playwright/cli@latest to download and run the Playwright testing utility at runtime.
[CREDENTIALS_UNSAFE]: The workflow requires reading sensitive information from local environment files (.env.local, .env.test) and browser profiles to manage authentication during testing.
[REMOTE_CODE_EXECUTION]: By executing the Playwright CLI via npx, the skill runs remote code from the npm registry. While the provider is well-known, it constitutes a runtime code download and execution.
[PROMPT_INJECTION]: The skill possesses a significant attack surface by ingesting data from a live web application (snapshots, console logs, network requests). A malicious application could inject instructions into these data streams to influence the agent's high-privilege operations.
Ingestion points: Browser console output, network request logs, and DOM snapshots collected via Playwright.
Boundary markers: None identified; the agent is instructed to directly evaluate and react to the browser state.
Capability inventory: Includes source code modification, shell command execution, and SQL query execution via Supabase MCP.
Sanitization: No specific sanitization or filtering logic is described for the ingested browser data.

test-playwright