The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for building plugins that process content from an Obsidian vault, creating a surface for potential injection from untrusted notes.\n
Ingestion points: As detailed in references/vault-and-editor.md, the skill instructs on using the Vault API (read, process) and Markdown post-processors to read and render file content.\n
Boundary markers: The instructions do not specify techniques for delimiting data from commands in the processed content.\n
Capability inventory: The skill documents capabilities to modify vault files, manipulate the active editor, and register global commands as seen in references/vault-and-editor.md and references/ui.md.\n
Sanitization: The skill mitigates risks by explicitly directing developers to avoid innerHTML/outerHTML and instead use safe DOM construction methods like createEl() and createDiv(), and to normalizePath() all user-supplied paths.\n- [COMMAND_EXECUTION]: The references/plugin-basics.md file contains setup instructions that involve shell commands such as mkdir, cd, and npm run dev for project initialization and development.\n- [EXTERNAL_DOWNLOADS]: The skill references external resources for development.\n
It provides instructions to clone the sample plugin from the official obsidianmd GitHub organization.\n
It identifies the obsidian npm package as the primary API dependency for plugins.

obsidian-dev-rules