obsidian-nvim
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates agent interaction with an Obsidian vault, which consists of user-provided markdown files. These files represent an ingestion surface for untrusted data that may contain malicious instructions intended to influence the agent's behavior.
- Ingestion points: Obsidian vault directories and markdown files (e.g.,
~/vaults/personal,~/vaults/work,.mdfiles) referenced inSKILL.mdandreferences/configuration.md. - Boundary markers: None identified. The skill does not provide instructions for the agent to ignore or delimit embedded instructions within vault content.
- Capability inventory: The skill documents capabilities for searching vault content (
:Obsidian search), creating/renaming notes (:Obsidian new,:Obsidian rename), inserting templates (:Obsidian template), and performing clipboard operations (:Obsidian paste_img). - Sanitization: No specific sanitization or filtering logic for note content is described.
- [EXTERNAL_DOWNLOADS]: The skill references several external Neovim plugins and system tools required for functionality. While these are standard in the Neovim community, they involve fetching code and binaries from external sources.
- Evidence:
obsidian-nvim/obsidian.nvim(Main plugin repository)nvim-lua/plenary.nvim,nvim-telescope/telescope.nvim,hrsh7th/nvim-cmp,saghen/blink.cmp(Plugin dependencies and integrations)- System tool requirements:
ripgrep,pngpaste,xclip,wl-clipboard.
Audit Metadata