The Agent Skills Directory

Command Execution (SAFE): The skill utilizes local commands such as mkdir and openspec to set up the development environment. These operations are restricted to the current project directory and align with the skill's stated purpose.
Dependency Analysis (SAFE): The skill requires the openspec CLI as a precondition. It does not attempt to download or install this dependency at runtime, instead instructing the user to install it manually if missing.
Indirect Prompt Injection (INFO): The skill processes JSON output from the openspec schema validate command. While this represents a surface for tool output poisoning if the project files are malicious, the impact is negligible as the agent is directed to display the information to the user for manual review. Ingestion Point: openspec schema validate JSON output; Boundary Markers: Absent; Capability Inventory: mkdir, openspec CLI; Sanitization: Display to user for manual inspection.

superspec-init