setup-vault
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a legitimate configuration tool that scans local vault directories (
.obsidian/,AGENTS.md,CLAUDE.md) to establish writing conventions. It does not perform network operations or interact with remote services.- [SAFE]: The tool incorporates a human-in-the-loop security measure by presenting a draft of the configuration file for user review and approval before any filesystem write occurs.- [SAFE]: No signs of obfuscation, credential theft, persistence, or privilege escalation were found. The skill's behavior is consistent with its stated purpose of vault organization.- [SAFE]: While the skill ingests user-provided content from the vault to determine style, the lack of high-risk capabilities (such as shell execution or network calls) and the mandatory approval step minimize the risk of indirect prompt injection. Findings for the indirect injection surface include: - Ingestion points:
AGENTS.md,CLAUDE.md, and note corpus folders (e.g.,Literature,Evergreen). - Boundary markers: Absent.
- Capability inventory: Local file writing (
.agents/slipbox/vault-conventions.md). - Sanitization: Absent.
Audit Metadata