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