The Agent Skills Directory

[SAFE]: The skill implements a robust workflow for requirements engineering. It operates purely on provided text inputs and produces structured markdown outputs using predefined templates (output-template.md, evidence-block.md).
[SAFE]: File system interactions are limited to reading upstream documents and writing results within a project workspace (e.g., 03-scenarios.md, status.json). No access to sensitive directories, environment variables, or user credentials was found.
[SAFE]: The skill includes a comprehensive 'Research Policy' that mandates a research plan before using web search tools and requires all findings to be documented with credibility ratings, which promotes safety and accountability.
[SAFE]: The configuration explicitly disables model invocation (disable-model-invocation: true), which prevents the skill from triggering unauthorized external AI model calls or recursive agent behavior.
[SAFE]: No obfuscation, prompt injection attempts, or persistence mechanisms were identified in the instructions or metadata.

scenario-restore