find-skills
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- Command Execution: The skill utilizes the
npx skillsCLI to perform search and management tasks. This involves executing subprocesses to extend the agent's functional environment. - External Downloads and Installation: It provides instructions to download and install modular packages from the
skills.shplatform and GitHub. These actions are fundamental to the skill's utility as an installer. - Indirect Prompt Injection Surface: The skill interpolates user-provided search terms into shell commands.
- Ingestion points: User queries are passed directly to the
npx skills find [query]command. - Boundary markers: The instructions do not define specific delimiters to isolate user input within the command string.
- Capability inventory: The skill can execute CLI commands and install external code via
npx. - Sanitization: There is no explicit mention of sanitizing or escaping user input before it is used as a command argument.
- Automated Execution Flags: The use of the
-yflag in thenpx skills addcommand automates the installation process by skipping manual confirmations, which is standard for streamlined package management.
Audit Metadata