The Agent Skills Directory

[SAFE]: The skill implements legal assessment logic through local Python scripts that utilize the standard library for JSON handling and text matching.
[SAFE]: No network activity, remote code execution, or unauthorized file access was detected. The scripts operate exclusively on local data provided by the user.
[SAFE]: Analysis of instructions and code confirms the absence of prompt injection, obfuscation, or persistence mechanisms.
[PROMPT_INJECTION]: The skill processes external text for threshold evaluation, representing an indirect prompt injection surface.
Ingestion points: scripts/dpia_threshold_checker.py via activity descriptions.
Boundary markers: Not present.
Capability inventory: Limited to local file R/W; no network or subprocess capabilities.
Sanitization: None. Risk is negligible due to the absence of execution paths for the ingested data.

dpia-assessment