The Agent Skills Directory

[SAFE]: No malicious patterns or unauthorized activities were detected. The skill is designed to assist in agent development and evaluation workflows.- [COMMAND_EXECUTION]: The skill utilizes local CLI tools, specifically 'kensa' and 'uv', to analyze execution traces and run agent evaluations. These commands are integral to the skill's documented functionality and are executed within the user's local environment.- [PROMPT_INJECTION]: The skill instructions provide proactive security guidance by warning against manual quoting of input placeholders and recommending the use of the platform's built-in sanitization (shlex.quote) to prevent command injection vulnerabilities in generated scenario files.- [SAFE]: A review of Category 8 (Indirect Prompt Injection) surfaces indicates that while the skill ingests project data (agents, scenarios, and traces), it does so to inform the design of structured YAML files rather than for direct prompt interpolation. The recommended use of shlex.quote for dynamic inputs further secures the execution path.

generate-scenarios