prototype
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to integrate with the host project's existing task runners (e.g.,
package.jsonscripts,Makefile,justfile) to facilitate the execution of prototypes. This is a standard and expected behavior for developer-oriented tools. - [DATA_EXFILTRATION]: Guidelines advise the agent to use in-memory state or local scratch files for prototypes to avoid unintended persistence or interference with production data systems.
- [PROMPT_INJECTION]: The skill does not contain instructions that attempt to bypass safety filters, override system prompts, or extract internal configuration. The instructional language is focused on task execution within the defined scope of prototyping.
Audit Metadata