The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install external packages from third-party sources.
Evidence: npm install -g @bitbonsai/mcpvault and cli-hub install obsidian in SKILL.md.
[COMMAND_EXECUTION]: The skill utilizes shell commands to install software and configure the agent's environment.
Evidence: Use of npm, npx, and codex mcp add commands in SKILL.md.
[DATA_EXFILTRATION]: The skill is designed to grant the agent read and write access to a local Obsidian vault, which contains private user data.
Evidence: Instructions for authorizing vault folders and providing vault paths to the MCP server.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads untrusted data from the user's Obsidian vault.
Ingestion points: Local files within the Obsidian vault path provided by the user (SKILL.md).
Boundary markers: None identified in the skill instructions to separate file content from system instructions.
Capability inventory: The skill provides read and write access to the local file system via the configured MCP server (SKILL.md).
Sanitization: No sanitization or validation of the vault content is mentioned before it is processed by the agent.

codex-obsidian