The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell commands to manage the testing environment, including starting development servers (e.g., npm run dev, ng serve) and running a local Node.js script (uat-auth-setup.mjs) to handle authentication state.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection due to its automated ingestion of external data. \n
Ingestion points: Reads specification files from the repository, analyzes branch changes via git diff, fetches pull request details via gh pr view, and inspects live website content through Playwright's browser_snapshot. \n
Boundary markers: The instructions do not define clear delimiters or "ignore instructions" markers when processing this external content. \n
Capability inventory: The agent has the capability to execute local Node.js scripts (in SKILL.md), write files to the local system for test cases and reporting (in SKILL.md), and execute arbitrary JavaScript in the browser (in references/agent-guide.md). \n
Sanitization: There is no explicit mechanism for sanitizing or validating the untrusted content read from specifications or browser snapshots before it is used to influence agent decisions.
[DYNAMIC_EXECUTION]: The skill utilizes the browser_evaluate tool to execute arbitrary JavaScript within the target application's browser context, which is a powerful capability required for testing but susceptible to misuse if the agent's instructions are compromised by external data.

uat-testing