skills/olafgeibig/skills/vault-ops/Gen Agent Trust Hub

vault-ops

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill has a significant indirect prompt injection surface. It explicitly instructs the agent to read and prioritize instructions from an AGENTS.md file found at the root of any selected vault. This allows untrusted data within a vault to override the agent's instructions. Evidence: Ingestion points include AGENTS.md and referenced files (SKILL.md); boundary markers are absent; the agent possesses file read/write and shell execution capabilities; no sanitization is specified for vault content.
  • [COMMAND_EXECUTION]: The skill relies on the execution of local shell commands. It provides templates for using ripgrep (rg) to search through notes, frontmatter, and links.
  • [DATA_EXFILTRATION]: The skill accesses local configuration files to discover vault paths. It reads ~/.vault-ops.json and Obsidian-specific application configuration files (e.g., obsidian.json) across macOS, Windows, and Linux. While this is the intended purpose of the skill, it involves accessing system-level application data.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 11:19 PM