gwb-zusammenschluss-anmeldung

Pass

Audited by Gen Agent Trust Hub on Jun 28, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No attempts to override system instructions or safety filters were detected. The instructions are focused on legal procedural steps.
  • [DATA_EXFILTRATION]: No unauthorized network activity or hardcoded credentials. External URLs point to official legal repositories including gesetze-im-internet.de, eur-lex.europa.eu, and bundeskartellamt.de.
  • [REMOTE_CODE_EXECUTION]: The skill does not perform remote code execution or download external scripts at runtime.
  • [OBFUSCATION]: No hidden text, encoded strings (Base64), zero-width characters, or homoglyph attacks were identified.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied transaction data and financial figures to generate a legal memorandum. While this is a data ingestion surface, there are no dangerous capabilities (like shell access or arbitrary network sends) for an attacker to leverage via injected text. The output is a structured text document.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 28, 2026, 08:40 AM
Security Audit — agent-trust-hub — gwb-zusammenschluss-anmeldung