obsidian-hardened

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill manages Obsidian vaults by generating commands for the obsidian-cli utility, covering actions like searching, creating, moving, and deleting Markdown notes.
  • [COMMAND_EXECUTION]: It implements defense-in-depth by explicitly mandating that the agent quote all user-provided strings in CLI arguments to prevent shell command injection.
  • [DATA_EXFILTRATION]: Includes a strict prohibition against piping any vault data (note content, search snippets, or paths) to network-transmitting commands such as curl, wget, or nc, ensuring notes remain local.
  • [EXTERNAL_DOWNLOADS]: Automates the installation of the necessary obsidian-cli binary via the Homebrew package manager using a third-party tap.
  • [SAFE]: Enforces data integrity through guardrails that prefer append-only editing over overwriting and require explicit user confirmation before performing destructive bulk operations.
  • [SAFE]: Adheres to privacy-by-design principles by instructing the agent to show only note titles by default in search results, hiding potentially sensitive snippets unless specifically requested.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 06:31 PM
Security Audit — agent-trust-hub — obsidian-hardened