The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides a run-code command that evaluates arbitrary JavaScript strings within the browser context. This allows the agent to execute custom logic, interact with internal page state, and manipulate browser permissions at runtime. Evidence is found in SKILL.md and references/running-code.md describing the usage of playwright-cli run-code with function strings.
[DATA_EXFILTRATION]: The skill facilitates the extraction of sensitive browser information, including authentication cookies, localStorage, and sessionStorage. It also provides commands to save the entire browser state to disk. While necessary for automation, these commands allow the retrieval of active session tokens. Evidence is found in SKILL.md and references/storage-state.md documenting commands like cookie-get, localstorage-list, and state-save.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the web (e.g., through snapshots, element attributes, and page text) which is then used by the agent to make decisions. This creates a vulnerability where malicious web pages could influence the agent's behavior.
Ingestion points: Browser snapshots, element text content (via eval), page content, and console logs are read into the agent's context in SKILL.md and references/test-generation.md.
Boundary markers: The instructions do not define explicit boundary markers or warnings to the agent to disregard instructions found within web page content.
Capability inventory: The agent can execute browser JavaScript (run-code), write files (state-save, screenshot, pdf), and intercept network requests (route) as documented throughout the skill.
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from browser interactions before it is presented to the agent.
[COMMAND_EXECUTION]: The skill uses command-line interfaces to execute playwright-cli, npx, and npm commands, allowing for the execution of browser automation tasks and the management of Playwright tests as detailed in SKILL.md and references/playwright-tests.md.
[EXTERNAL_DOWNLOADS]: The skill instructions suggest the global installation of the Playwright CLI from the official npm registry (npm install -g @playwright/cli@latest). This is a reference to a well-known service and organization.

playwright-cli