obsidian-wiki
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill makes extensive use of an external
obsidianCLI tool to perform file operations including reading, creating, appending, and searching within the Obsidian vault, as well as retrieving vault metadata. This functionality is essential for its role as a knowledge base manager. - [EXTERNAL_DOWNLOADS]: The skill performs automated web searches to gather information on concepts and check for content updates. This involves fetching data from external, untrusted sources which is then incorporated into the local knowledge base.
- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection (Category 8) because it ingests and processes untrusted data from both web searches and existing user notes.
- Ingestion points: External web search results used for 'deep writing' concepts and local articles read via the
obsidian readcommand. - Boundary markers: There are no specified delimiters or instructions to ignore embedded commands within the fetched external content or user notes.
- Capability inventory: The skill possesses file system write permissions (
create,append,writetool), file read permissions, web search access, and the ability to delegate tasks to subagents. - Sanitization: The instructions do not describe any sanitization, filtering, or validation steps for the content retrieved from the web before it is processed by the AI or written to the vault.
Audit Metadata