semantic-safe-implementation-refactor
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is composed entirely of instructional Markdown documents and templates that establish a workflow for software refactoring. It does not perform any automated operations outside of providing instructions to the agent.
- [NO_CODE]: No executable code, shell scripts, or binaries are included in the skill package. It leverages the agent's native ability to read and edit code within the defined process.
- [DATA_EXFILTRATION]: There are no commands or instructions that involve network activity, data exfiltration, or access to sensitive local environment variables or credential files.
- [PROMPT_INJECTION]: The instructions do not use any patterns aimed at overriding the agent's system prompt, safety filters, or core behavior. The language is focused on engineering best practices and preserving existing software semantics.
Audit Metadata