software-foerderung-rd-tax

Pass

Audited by Gen Agent Trust Hub on Sep 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to perform live verification of legal norms and case law using official and well-known German legal databases including gesetze-im-internet.de, dejure.org, openJur, and the databases of the BVerfG, BGH, and EuGH.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external legal documents and case facts which may originate from untrusted sources or legal adversaries.
  • Ingestion points: The agent is instructed in SKILL.md to collect and review various document types including contracts, pleadings, and expert opinions from external parties.
  • Boundary markers: No explicit delimiters are provided in SKILL.md to distinguish between instruction context and external data.
  • Capability inventory: Based on SKILL.md, the skill is restricted to analytical tasks and document generation without access to high-risk system tools.
  • Sanitization: SKILL.md lacks specific procedures for sanitizing or validating external document text.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 30, 2026, 01:26 PM