The Agent Skills Directory

[SAFE]: The skill serves as a documentation overlay for Claude Code conventions, defining standards for plugin manifests, agent frontmatter, and hook event schemas.
[DYNAMIC_CONTEXT_INJECTION]: The skill documents the !command syntax used in Claude Code for dynamic context injection. As this is part of a documentation guide and does not include malicious commands meant for execution during the loading of this specific skill, it is considered safe practice.
[EXTERNAL_DOWNLOADS]: The skill references authoritative documentation at code.claude.com. These are links to well-known service documentation and do not constitute a security risk.
[COMMAND_EXECUTION]: The skill describes how commands, hooks, and LSP servers are configured in the Claude Code environment. It does not contain any immediate command execution payloads.

conventions-claude