bizzon

Pass

Audited by Gen Agent Trust Hub on Apr 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the membrane CLI for core integration tasks such as authentication and API interaction in SKILL.md.
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing the @membranehq/cli package from the official NPM registry to enable interaction with the Membrane platform.
  • [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection via data retrieved from the Bizzon API. 1. Ingestion points: Data enters the context through membrane action run and membrane request commands in SKILL.md. 2. Boundary markers: Boundary markers are absent; the skill does not define specific delimiters to separate external data from system instructions. 3. Capability inventory: The agent can execute shell commands using the membrane CLI tool as documented in SKILL.md. 4. Sanitization: Sanitization is absent; no validation or filtering of the external API responses is described.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 22, 2026, 05:10 PM