The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION] (LOW): The skill is susceptible to indirect prompt injection through its task handling logic.
Ingestion points: The $TASK environment variable is used within the pre hook string.
Boundary markers: Absent; the variable is interpolated directly into a shell command string.
Capability inventory: The skill can execute shell commands via npm test and shell hooks.
Sanitization: Absent; there is no evidence of escaping or validation for the $TASK input before it is used in the echo command.
[COMMAND_EXECUTION] (SAFE): The execution of npm test is a primary function of the skill and is appropriate for a TDD-focused agent.
[BEST_PRACTICE] (SAFE): The pre hook contains a typo ($dev$null instead of /dev/null), which results in redirection to a file named null if the $dev variable is empty. This is a functional bug rather than a direct security threat.

agent-tdd-london-swarm