NYC

axiom-ios-vision

Pass

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: LOW
Full Analysis
  • Prompt Injection (SAFE): The skill contains no instructions designed to bypass safety filters or override system prompts. It uses standard instructional formatting to define routing logic.
  • Data Exposure & Exfiltration (SAFE): No hardcoded credentials, sensitive file paths, or network operations were found in the file.
  • Obfuscation (SAFE): No encoded content, zero-width characters, or homoglyphs were detected.
  • Unverifiable Dependencies & Remote Code Execution (SAFE): The skill does not reference external packages or remote scripts. It only references other internal skills (e.g., axiom-vision) via semantic routing.
  • Privilege Escalation (SAFE): No commands related to elevated permissions or system-level modifications are present.
  • Persistence Mechanisms (SAFE): The skill does not attempt to modify shell profiles, cron jobs, or startup services.
  • Metadata Poisoning (SAFE): Metadata fields (name, description, license) accurately reflect the purpose of the skill.
  • Indirect Prompt Injection (SAFE): As a routing-only skill, it does not directly ingest or process external untrusted data. It merely directs the agent on which tool to use for specific intents.
  • Time-Delayed / Conditional Attacks (SAFE): No logic gating behavior based on time or specific environment variables was found.
  • Dynamic Execution (SAFE): The skill contains no code generation, runtime compilation, or unsafe deserialization patterns.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 15, 2026, 10:31 PM