vault-insights

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes the entire contents of a local Obsidian vault to build mental models and extract insights. This creates an indirect prompt injection surface where malicious content within notes could attempt to manipulate the agent's logic.
  • Ingestion points: Markdown files (*.md) in the '~/Documents/obsidian-vault/' directory.
  • Boundary markers: None identified; the instructions direct the agent to 'read every note' without delimiters or instructions to ignore embedded commands.
  • Capability inventory: The skill uses the 'Edit' tool (writing to daily notes), the 'Bash' tool (executing CLI commands), and the 'Task' tool (subagent orchestration).
  • Sanitization: There is no mention of sanitizing, escaping, or validating the note content before it is processed.
  • [COMMAND_EXECUTION]: The skill uses the 'Bash' tool to invoke the GitHub CLI ('gh pr view') using PR identifiers found within note content. While this targets a well-known service, the execution parameters are derived from untrusted local data.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 07:07 PM