The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands using the qmd utility (or a path specified by QMD_CLI) to update, embed, and verify vector search indexes for the Obsidian vault. These commands are run after vault write operations to ensure the index remains current.
[INDIRECT_PROMPT_INJECTION]: The skill reads external data from the vault, such as index.md, to determine existing categories and pages. This information is used to influence the agent's logic when drafting new dashboard views, creating a potential surface for indirect prompt injection if the vault content is malicious.
Ingestion points: Accesses $OBSIDIAN_VAULT_PATH/index.md and other vault files to understand content structure.
Boundary markers: No specific delimiters or instructions to ignore embedded commands within the vault data are provided.
Capability inventory: The skill has the ability to write new .base and .md files, append to a log.md file, and execute local shell commands via the qmd CLI.
Sanitization: The skill does not implement explicit sanitization or validation of the ingested vault content before using it to generate dashboard configurations.

wiki-dashboard