agent-eval
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill contains no executable scripts, binaries, or automated shell commands. It functions entirely as a documentation guide for using an external CLI tool.
- [EXTERNAL_DOWNLOADS]: The skill references an external GitHub repository for the
agent-evaltool (github.com/joaquinhuigomez/agent-eval). It explicitly instructs the user to review the source code manually before installation, following security best practices. - [COMMAND_EXECUTION]: While the documentation includes example shell commands (e.g.,
mkdir,agent-eval run), these are provided as usage instructions for the user to execute manually rather than being automated by the skill. - [INDIRECT_PROMPT_INJECTION]: The skill describes a system for processing tasks defined in YAML files, which is a potential surface for indirect injection. However, since the skill itself provides no code to automate this process, it currently presents no active risk.
Audit Metadata