The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the target application's user interface and API responses, which are then used to guide the behavior of sub-agents. \n
Ingestion points: projects/<name>/qa/README.md (which may contain user-provided instructions), Playwright browser content, and API response bodies from curl. \n
Boundary markers: The skill does not define specific delimiters or 'ignore' instructions to separate ingested product data from its own core logic. \n
Capability inventory: Spawns specialized agents, writes markdown files to the repository (projects/<name>/qa/qa-run-*.md), creates backlog tickets, and executes network requests via curl and Playwright. \n
Sanitization: There is no evidence of filtering or validation for the content retrieved from the live product before it influences agent actions.\n- [COMMAND_EXECUTION]: The skill utilizes curl and Playwright MCP servers to interact with the environment, involving the execution of shell-level operations for testing and data retrieval.\n- [DATA_EXFILTRATION]: The skill performs network operations via curl to reach the application's API and database stores. Although intended for observability, these operations target endpoints that are not contained within a predefined safety whitelist.

qa-sweep