skills/klotzkette/claude-fuer-deutsches-recht/master-franchise-und-grenzueberschreitende-expansion/Gen Agent Trust Hub
master-franchise-und-grenzueberschreitende-expansion
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill contains standard instructional language for a legal AI agent. There are no attempts to override safety filters, bypass constraints, or elicit system prompts.
- [DATA_EXPOSURE]: No hardcoded credentials, API keys, or sensitive file paths were detected. The instructions for collecting client documents (contracts, emails) are standard procedure for legal analysis and do not involve exfiltration mechanisms.
- [OBFUSCATION]: The content is written in clear, plain German. No base64 encoding, zero-width characters, homoglyphs, or other obfuscation techniques were found.
- [EXTERNAL_DOWNLOADS]: The skill references well-known German legal databases (gesetze-im-internet.de, dejure.org, openJur) for verifying legal norms. These are trusted reference sources and do not involve downloading or executing untrusted binaries.
- [REMOTE_CODE_EXECUTION]: There are no patterns involving remote script execution, package installations, or dynamic code execution.
- [INDIRECT_PROMPT_INJECTION]: While the skill involves processing external data (contracts, audit reports), it does not provide any exploitable tools or scripts that could be targeted by embedded instructions in those documents. The risk is inherent to all document-processing agents but remains safe in this context.
- [DYNAMIC_CONTEXT_INJECTION]: No use of the dynamic exclamation-backtick syntax was found in the SKILL.md file.
Audit Metadata