material-contract-schedule
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection Surface: The skill is designed to ingest and act upon text from external sources, specifically purchase agreement drafts and diligence findings. This introduces a potential consideration where instructions embedded within those documents might influence agent behavior.
- Evidence Chain for Data Processing: (1) Ingestion points: External documents such as purchase agreement drafts and diligence findings are loaded at runtime (SKILL.md). (2) Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the ingested text to prevent instruction leakage. (3) Capability inventory: The skill performs file reads and writes within specific matter-related directories (~/.claude/plugins/config/claude-for-legal/) to maintain context and overrides. (4) Sanitization: No specific sanitization or validation steps for external content are defined in the workflow.
Audit Metadata