The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to perform live verification of legal norms and judicial rulings using external legal information portals.
Evidence: The instructions explicitly direct the agent to use gesetze-im-internet.de, dejure.org, eur-lex.europa.eu, and openJur. These are well-known and official legal repositories, and their use for verifying source material is a standard, safe operation for this use-case.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as contracts and legal documents, which constitutes a potential injection surface.
Ingestion points: As described in SKILL.md under the 'Arbeitsweg' and 'Sofortstart' sections, the skill ingests user-provided files including administrative acts, contract documents, expert reports, and code files.
Boundary markers: There are no specific instructions or delimiters defined to separate user data from the agent's system instructions or to warn the agent about potential embedded commands in those documents.
Capability inventory: The skill is designed for analytical and text-generation tasks (producing risk matrices, memos, and legal drafts). It does not exhibit dangerous capabilities such as arbitrary command execution, file system modification, or unauthorized network transmissions.
Sanitization: No explicit input sanitization or validation protocols are described in the instructions.

software-69a-computerprogramm