The Agent Skills Directory

[SAFE]: Analysis of the skill instructions and schema reveals no malicious intent. The skill is strictly for configuring a specific Obsidian plugin using YAML files.
[INDIRECT_PROMPT_INJECTION]: The skill identifies a data ingestion surface by allowing the agent to process Obsidian note metadata and properties.
Ingestion points: File metadata (e.g., file.tags, file.path, file.mtime) and note frontmatter properties defined in the properties and filters sections of SKILL.md.
Boundary markers: Data is ingested through structured YAML configuration files, which provide inherent delimitation.
Capability inventory: The skill is limited to creating and modifying .base files; it does not invoke high-risk tools, shell commands, or network operations based on the ingested data.
Sanitization: Relies on the underlying Obsidian Bases plugin to safely evaluate formula expressions and filters.

obsidian-bases