json-canvas
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a documentation resource for the JSON Canvas specification, providing standard schemas and validation steps for visual note-taking data.
- [PROMPT_INJECTION]: The skill handles .canvas files which ingest untrusted text data. This creates a surface for indirect prompt injection where malicious instructions could be embedded in text nodes. However, the risk is inherent to the data format and mitigated by the skill's focus on structural JSON manipulation.
- Ingestion points: Workflows 2 and 4 describe reading and parsing existing .canvas files from the filesystem.
- Boundary markers: Instructions emphasize JSON schema validation and ID uniqueness but do not specify content sanitization for text nodes.
- Capability inventory: The skill utilizes file system write capabilities to update .canvas files.
- Sanitization: Validation is limited to JSON structure and mandatory field presence.
Audit Metadata