The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands using npx to run the test suite (npx shiplight test) and to update its own dependencies (npx -y skills@latest update -y).
[REMOTE_CODE_EXECUTION]: The skill uses npx -y skills@latest update -y to download and execute code from the npm registry. This is part of the vendor's self-update mechanism but represents a remote code execution vector.
[EXTERNAL_DOWNLOADS]: The skill periodically checks for updates and downloads the latest version of the skills package from a remote repository via npx.
[CREDENTIALS_UNSAFE]: The skill accesses and utilizes "storage state files," which are used by Playwright to store sensitive session information such as cookies and authentication tokens.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the application under test.
Ingestion points: The agent reads the application's DOM through inspect_page and analyzes browser logs (console and network) to diagnose failures.
Boundary markers: The instructions do not provide explicit delimiters or warnings to ignore instructions embedded within the application's UI or logs.
Capability inventory: The agent can execute shell commands (npx), modify local YAML files, and interact with the browser via the act tool.
Sanitization: There is no documented mechanism for filtering or sanitizing the content retrieved from the web page before it is processed by the agent to generate new test steps.

triage