tabular-review-lawvable

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFE
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill processes content from untrusted external documents (PDF, DOCX) which could potentially contain malicious instructions intended to mislead the LLM sub-agents during the extraction process.
  • Ingestion points: Local PDF and DOCX files identified via user-provided folder paths in Step 2.
  • Boundary markers: Absent; the sub-agent prompt template does not use specific delimiters or instructions to ignore embedded commands within the document text.
  • Capability inventory: Reading local filesystem, launching background sub-agents (Task), and writing Excel files (xlsx skill).
  • Sanitization: None; the raw text of documents is processed by the sub-agents for extraction.
  • [Data Exposure] (SAFE): The skill generates Excel files containing absolute file paths in file:// hyperlinks. While this exposes local path structures, it is a core feature of the tool's document review matrix purpose and is initiated only via explicit user-provided paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 04:45 AM