The Agent Skills Directory

[PROMPT_INJECTION]: The skill defines syntax for embedding content from external files within a vault, which creates an ingestion point for potentially untrusted data that could be used for indirect prompt injection.
Ingestion points: The main skill instructions in SKILL.md and the reference file references/EMBEDS.md detail how to use the ![[embed]] syntax to pull content into the current context.
Boundary markers: The instructions do not define delimiters or provide guidance to the agent on how to distinguish between passive note content and active instructions when processing embedded files.
Capability inventory: While the skill itself is markdown-only, it is designed for use by agents that have file system access to read and write notes within a vault.
Sanitization: No methods for sanitizing, escaping, or validating the content of embedded files are mentioned in the instructions.

obsidian-markdown