json-canvas

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it instructs the agent to read and process content from .canvas files which can contain attacker-controlled text.
  • Ingestion points: The agent parses node data from .canvas files, including user-provided markdown in text nodes.
  • Boundary markers: There are no instructions to use delimiters or ignore embedded instructions when reading node content.
  • Capability inventory: The skill facilitates file system read and write operations to manage canvas files.
  • Sanitization: Validation is restricted to JSON schema adherence and reference integrity, with no sanitization of the actual text content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 08:19 AM
Security Audit — agent-trust-hub — json-canvas