skills/ravi-hq/deepthink-skills/bezos/Gen Agent Trust Hub

bezos

Pass

Audited by Gen Agent Trust Hub on Apr 23, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it incorporates arbitrary user input directly into the prompts for five specialized sub-agents.
  • Ingestion points: User-provided business ideas, decisions, or diagnoses entered via the /bezos command are passed as the primary analysis subject to five agents.
  • Boundary markers: The instructions use simple placeholders like [INSERT FULL IDEA/DECISION TEXT] without robust delimiters, XML tags, or instructions for the sub-agents to ignore potentially malicious commands embedded in the input.
  • Capability inventory: The skill's environment grants access to tools including Bash, Write, Read, Glob, and WebFetch, which could be exploited if a sub-agent is manipulated by adversarial input.
  • Sanitization: There is no logic or instruction provided to sanitize, escape, or validate the user-supplied input before it is processed by the sub-agents.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 23, 2026, 11:52 PM