json-canvas
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill identifies an attack surface for indirect prompt injection because it is designed to read, parse, and modify data that can contain untrusted natural language instructions.
- Ingestion points: The common workflows described in SKILL.md (Add a Node, Edit an Existing Canvas) require the agent to read and parse existing .canvas files from the filesystem.
- Boundary markers: There are no instructions for the agent to use delimiters or "ignore instructions" tags when processing the contents of text nodes in the JSON structure.
- Capability inventory: The skill allows the agent to read from and write to the local filesystem (creating or modifying .canvas files).
- Sanitization: The instructions do not specify any validation or sanitization for the 'text' attribute of nodes, which could be used to host malicious instructions intended for the agent or downstream skills like 'obsidian-markdown'.
Audit Metadata