reviewing-an-artifact
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill package is composed entirely of Markdown instructions and documentation. No executable scripts, binaries, or active code components are present.
- [EXTERNAL_DOWNLOADS]: The skill includes links to official Anthropic documentation and technical blog posts regarding best practices for agent skills. These references are to trusted sources and do not involve the download or execution of external code.
- [DATA_EXFILTRATION]: There are no commands or instructions that suggest unauthorized file access or data exfiltration. The skill's operations are limited to reading its own internal reference files and the artifact being reviewed.
- [PROMPT_INJECTION]: The instructional content does not contain attempts to override agent behavior, bypass safety guardrails, or leak system prompts. It maintains a strictly professional role as a code review tool.
- [INDIRECT_PROMPT_INJECTION]: Although the skill is designed to process external artifacts, it possesses no high-privilege capabilities (such as shell execution, network writes, or file modification) that could be exploited by an injection attack within the reviewed content.
Audit Metadata