The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process natural language acceptance criteria and QA flows to generate executable test steps and ad-hoc browser automation commands via the ta run utility. This creates a vulnerability surface where malicious instructions embedded in the input data could influence the generated tests or browser actions.
Ingestion points: The agent is instructed to ingest user-provided acceptance criteria, user journeys, and QA instructions from the project context.
Boundary markers: The skill does not provide specific instructions or delimiters to separate untrusted user data from the agent's logic, increasing the risk of instruction confusion.
Capability inventory: The skill allows the agent to execute the ta CLI for test creation, management of project credentials, and performing local or remote test runs.
Sanitization: No instructions are provided for sanitizing or validating user-provided strings before they are incorporated into CLI command arguments or JSON payloads for test creation.
[COMMAND_EXECUTION]: The skill enables the agent to execute the ta (TesterArmy) CLI tool. This grants the agent significant control over the QA environment, including the ability to manage dashboard resources, handle sensitive credentials for login flows, and execute browser-based tests locally or in the cloud.

testerarmy-cli