find-skills
Warn
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill facilitates the retrieval and installation of agent extensions from remote repositories. It references public repositories from organizations like Vercel Labs and Anthropic as examples of available skills.- [REMOTE_CODE_EXECUTION]: The skill uses the
npx skills addcommand to download and execute code from external sources to extend the agent's functionality.- [COMMAND_EXECUTION]: The skill executes theskillsCLI tool usingnpx, which involves downloading and running packages from the NPM registry.- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection via user-provided search terms. - Ingestion points: User queries are interpolated into
npx skills findcommands inSKILL.md. - Boundary markers: There are no delimiters or warnings to treat search results as untrusted data.
- Capability inventory: The skill can execute shell commands and install software packages.
- Sanitization: No sanitization or escaping of user input is performed before execution in the shell.- [PRIVILEGE_ESCALATION]: Instructions include the use of the
-yflag during installation, which suppresses user confirmation prompts and allows the agent to install external software autonomously.
Audit Metadata