obsidian-markdown
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external files and repository content, which creates a potential surface for indirect prompt injection. However, it mitigates this by instructing the agent to treat all imported content as data rather than instructions.
- Ingestion points: Local Obsidian markdown files, repositories, and external data sources accessed via SandBase.
- Boundary markers: Explicit instructions to treat imported content as untrusted data.
- Capability inventory: File reading and writing, and network operations through the SandBase API suite.
- Sanitization: Requirement to validate serialized files after editing and preserve existing schemas.
- [EXTERNAL_DOWNLOADS]: The skill utilizes the SandBase API suite (sandbase_discover, sandbase_inspect, sandbase_run, etc.) to fetch external content. The instructions enforce a strict confirmation workflow where the agent must present costs and parameters to the user before executing any paid or external calls.
Audit Metadata