The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses WebFetch to retrieve regulatory documents from authoritative institutional sources including eur-lex.europa.eu, curia.europa.eu, and edpb.europa.eu. These are well-known official domains and are documented as the primary sources for the skill's legal context.
[PROMPT_INJECTION]: While the skill processes untrusted user-provided content (PDFs, URLs, and text), it implements a comprehensive 'Analytical Canon' and 'Anti-hallucination protocol' in references/analytical_canon.md. These guidelines instruct the model to distinguish between source text and commentary, verify citations via live research, and flag uncertainties, effectively mitigating the risk of indirect prompt injection.
[COMMAND_EXECUTION]: The skill uses pdf-processing-anthropic and WebFetch for its core functions. These are standard platform tools. It does not attempt to execute arbitrary shell commands, escalate privileges, or modify system configurations.
[DATA_EXFILTRATION]: No evidence of unauthorized data transmission was found. The skill is designed to write its outputs (legal prompts and analyses) to the user's local working directory for review.

legal-analysis-forge-oliver-schmidt-prietz