The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install and execute code from unverified third-party sources to enable core functionality.
Recommends running uvx mcp-obsidian, which downloads and executes a package from a public registry.
Recommends running npx @bitbonsai/mcpvault@latest, which downloads and executes code from a public package manager.
[COMMAND_EXECUTION]: The skill relies on complex shell operations that include disabling standard security protocols.
Instructions recommend setting the environment variable NODE_TLS_REJECT_UNAUTHORIZED to 0. This disables TLS certificate validation for the entire Node.js process, which can expose the system to man-in-the-middle attacks if non-local traffic is involved.
Provides curl commands that handle sensitive API tokens and vault data through the terminal.
[PROMPT_INJECTION]: The design of the ingestion pipeline presents a risk of indirect prompt injection.
Ingestion points: The skill processes raw source documents stored in the .raw/ directory.
Boundary markers: There are no specified delimiters or system-level instructions to ignore malicious commands embedded within the source files.
Capability inventory: The skill utilizes Write, Edit, and Bash tools, providing a significant capability surface if malicious instructions in ingested data are followed.
Sanitization: The skill does not describe any validation or sanitization process for external content before it is incorporated into the knowledge base.

wiki