itr-saas-vertrag-bauleiter
Pass
Audited by Gen Agent Trust Hub on Aug 23, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill follows standard legal information processing workflows and does not contain any malicious code, hardcoded secrets, or unauthorized network activity. It references trusted German legal databases like gesetze-im-internet.de and dejure.org for normative verification.- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core function of processing external, untrusted legal documents.
- Ingestion points: The skill ingests user-provided files such as 'Verwaltungsakte' (administrative acts), 'Vertragsurkunden' (contract documents), 'Schriftsätze' (legal pleadings), and 'Protokolle' (minutes) as described in the documents and evidence collection sections.
- Boundary markers: There are no explicit boundary markers or instructions to the agent to disregard instructions potentially embedded within these documents.
- Capability inventory: The skill is capable of generating structured legal products, including memos, check-lists, and drafting structures for lawsuits or contracts. It lacks high-risk capabilities such as remote command execution or network-based data exfiltration.
- Sanitization: The skill does not define any sanitization or validation logic for the content processed from external documents.
Audit Metadata