The Agent Skills Directory

[DATA_EXFILTRATION]: The skill accesses a sensitive local configuration file located at ~/.test_infra_config.yaml. This file is required for Pulumi-based infrastructure provisioning as part of the skill's primary function.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute dda inv commands. It dynamically constructs these commands by interpolating variables derived from user input ($ARGUMENTS).
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (specifically command injection). It incorporates untrusted user input from $ARGUMENTS into shell commands without explicit instructions for sanitization or escaping.
Ingestion points: User-provided $ARGUMENTS parsed in the instructions.
Boundary markers: None specified for the shell command construction, though the skill requires manual user confirmation before execution.
Capability inventory: Use of the Bash tool to run constructed strings (SKILL.md).
Sanitization: The instructions lack specific validation or escaping requirements for input flags like --configparams or --targets.

run-e2e