cloud-convert

Pass

Audited by Gen Agent Trust Hub on Apr 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill operates by running membrane CLI commands for authentication, connection management, and file conversion tasks.\n- [EXTERNAL_DOWNLOADS]: The skill requires the @membranehq/cli tool, which is downloaded from the official NPM registry and provided by the skill author.\n- [PROMPT_INJECTION]: The skill's description erroneously includes CRM-related tasks (Deals, Leads), which is inconsistent with the Cloud Convert tool and could mislead the agent regarding its capabilities.\n- [PROMPT_INJECTION]: The skill processes files and URLs, creating an indirect prompt injection surface if the agent reads the output of the conversions.\n
  • Ingestion points: Actions like import-file-from-url and convert-file in SKILL.md.\n
  • Boundary markers: None identified to separate external file content from agent instructions.\n
  • Capability inventory: Execution of shell commands via the membrane CLI as described in SKILL.md.\n
  • Sanitization: Content sanitization for conversion outputs is not mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 30, 2026, 01:41 PM
Security Audit — agent-trust-hub — cloud-convert