Opposing Counsel Review
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill is comprised entirely of instructional markdown and YAML metadata. It does not contain any Python scripts, Node.js code, shell commands, or other executable components.
- [PROMPT_INJECTION]: The skill uses strong persona instructions to act as an adversarial counsel. These instructions are limited to legal reasoning and do not attempt to bypass core AI safety filters, extract system prompts, or override platform-level constraints.
- [SAFE]: No credential exposure, sensitive file access, or remote code downloads were identified. The skill operates solely within the text-processing context provided by the user.
Audit Metadata