The Agent Skills Directory

[SAFE]: The skill is dedicated to managing project documentation such as GLOSSARY.md and ADR files. It does not include any executable code or command scripts.
[INDIRECT_PROMPT_INJECTION]: The skill reads existing codebase files and user input to refine domain models. This creates a surface for indirect prompt injection; however, as the skill lacks high-privilege capabilities such as network access or arbitrary command execution, this surface does not present a security risk.
[SAFE]: The skill reads configuration files (AGENTS.md, CLAUDE.md) to determine the project's documentation root. This is a standard operation for repository-based tools and does not involve accessing sensitive system credentials.

domain-modeling