Skills
skills/dvdsgl/claude-canvas/document/Gen Agent Trust Hub

document

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted markdown content (e.g., from emails or web drafts) via the content parameter, which presents a surface for indirect prompt injection. Malicious instructions could be hidden in the text and interpreted by the agent during display or selection processing.
  • Ingestion points: content field in DocumentConfig within SKILL.md.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the configuration.
  • Capability inventory: Subprocess calls to src/cli.ts for rendering/spawning scenarios as documented in SKILL.md.
  • Sanitization: No input validation or sanitization of the markdown source is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 03:22 AM