The Agent Skills Directory

[COMMAND_EXECUTION]: The bash scripts scripts/generate_test_cases.sh and scripts/create_bug_report.sh use the eval command within the prompt_input function to assign user-supplied input to variable names. This implementation is insecure as it allows for command substitution (e.g., using $() or backticks) within the input, which can lead to arbitrary command execution in the context of the user's shell.
[PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection (Category 8) due to its workflow of processing external, untrusted content. Ingestion points: Feature requirements provided by users and Figma design specifications retrieved via MCP. Boundary markers: The skill lacks any explicit delimiters or instructions to the agent to ignore instructions embedded within the source requirements or Figma metadata. Capability inventory: The skill has the capability to generate and execute shell scripts through its provided automation tools. Sanitization: There is no evidence of input validation or escaping logic to prevent malicious payloads in the requirements from influencing the output generation or the behavior of the agent.

qa-test-planner