lenny-skills
Warn
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_CONTEXT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill is designed to fetch and install multiple external skill modules from a repository located at
https://github.com/RefoundAI/lenny-skills. It uses thenpx skills addcommand to pull these resources into the agent's local environment, representing an unverifiable package installation from a non-whitelisted source. - [COMMAND_EXECUTION]: The skill's installation logic relies on the execution of shell commands. This includes checking for existing files in the
~/.agents/skills/directory, utilizinggrepto verify file content, and runningnpxto manage dependencies. These commands are intended to be executed by the agent to automate the setup process. - [DYNAMIC_CONTEXT_INJECTION]: The documentation explicitly instructs the agent to use the
!prefix (e.g.,!command) to execute shell operations directly within the chat context. This mechanism is used to facilitate the installation of the sub-skills, creating a pathway where natural language instructions are translated into system-level command execution without explicit user isolation.
Audit Metadata