contract-redline

Pass

Audited by Gen Agent Trust Hub on Apr 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE]: The skill is composed entirely of markdown instructions and does not include any scripts, executables, or dependencies.
  • [SAFE]: Analysis of the instructions reveals no attempts at prompt injection, credential theft, or persistence. The skill performs intended business functions using provided context.
  • [PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by processing external contract documents. Ingestion points: Contract text (SKILL.md). Boundary markers: Absent. Capability inventory: Text generation only; no dangerous tool access detected. Sanitization: Absent. This surface is considered safe given the limited capability of the skill to perform actions beyond text output.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 17, 2026, 04:43 PM