The Agent Skills Directory

Command Execution: The skill utilizes the uv package manager to execute the design-agent CLI tool. This is a project-specific command used for listing and running evaluation tasks, which is the primary purpose of the skill.
Indirect Prompt Injection Surface: The skill instructs the agent to read and summarize evaluation reports (report.json, README.md). While these are generated by the evaluation framework, reading data produced by tests involves a theoretical surface for indirect instructions.
Ingestion points: Evaluation reports (report.json, README.md) and task output files located in workspace/out/evals/.
Boundary markers: None specified in the instructions.
Capability inventory: Execution of shell commands via uv run and file system read access.
Sanitization: No explicit sanitization or filtering of the report content is mentioned.

design-agent-evals