learning-vault

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands (mkdir, cp) to provision the Obsidian vault structure and install the Dataview plugin. These commands operate within the user's home directory (defaulting to ~/Brains/) to set up the environment.
  • [EXTERNAL_DOWNLOADS]: The skill bundles the Obsidian Dataview plugin (including main.js, manifest.json, and styles.css). While these files are local to the skill package, they introduce third-party executable JavaScript into the user's Obsidian workspace as part of the setup process.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its ability to process external URLs (e.g., course curricula) provided by the user to generate notes.
  • Ingestion points: User-provided URLs for courses and learning goals in SKILL.md.
  • Boundary markers: Absent; there are no specific instructions or delimiters used to separate the external curriculum data from the agent's instructions.
  • Capability inventory: The skill performs file creation and directory management within the local file system.
  • Sanitization: Absent; the skill does not describe any methods for validating or filtering content retrieved from the external URLs.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 10:20 PM
Security Audit — agent-trust-hub — learning-vault