The Agent Skills Directory

[SAFE]: The skill functions as a document generation assistant and does not utilize any restricted tools, network access, or shell execution capabilities. All processing is confined to text analysis and formatting.
[PROMPT_INJECTION]: The skill ingests untrusted user data in the form of case backgrounds, creating an indirect prompt injection surface. However, the lack of any actionable capabilities (such as file system writes or network calls) prevents this surface from being exploited for malicious purposes.
[DATA_EXFILTRATION]: No exfiltration risks were identified. The instructions include a mandatory 'anonymization check' that directs the agent to stop if real names or agencies are detected, which is a proactive measure for handling potentially sensitive legal information.

bail-application