analytics-setup-architect

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No prompt injection patterns detected. The instructions focus on planning and adherence to a safety contract.
  • [DATA_EXFILTRATION]: No exfiltration patterns found. The skill explicitly forbids storing analytics tokens, cookies, or secrets in its safety rules.
  • [COMMAND_EXECUTION]: The SKILL.md file contains a validation section with shell commands for linting and validation (e.g., using python3). These are standard development tools and do not involve untrusted input or malicious execution.
  • [REMOTE_CODE_EXECUTION]: No remote code execution or untrusted dependency downloads were identified.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user requirements to create plans. It mitigates risks by using a structured safety contract (references/analytics-safety-contract.md) and explicit safety rules that limit the agent's capabilities to planning rather than execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 07:30 AM
Security Audit — agent-trust-hub — analytics-setup-architect