The Agent Skills Directory

[DATA_EXFILTRATION]: The skill accesses local configuration files in ~/.claude/plugins/config/claude-for-legal/ to retrieve regulatory thresholds and AI system lists. It also writes generated reports to the same directory tree. These file operations are localized to the plugin workspace and are essential for its governance role.
[PROMPT_INJECTION]: The skill processes user-supplied AI system descriptions (ingestion point: SKILL.md) and interpolates them into assessment templates. It lacks explicit boundary markers or sanitization logic to separate this untrusted data from the prompt instructions. This creates an indirect prompt injection surface where the agent could be influenced to produce biased assessments, although such behavior is inherent to its primary function of report generation and no active malicious patterns were found.

aia-generation