find-skills
Warn
Audited by Gen Agent Trust Hub on Jun 28, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill facilitates downloading content from external sources including GitHub, GitLab, and other Git/HTTP endpoints via the
npx skills addcommand. - [REMOTE_CODE_EXECUTION]: The primary function of the skill is to install and enable external code within the agent environment. It references repositories from known providers like Vercel Labs, but also supports installation from any user-provided repository or local path, allowing for the execution of unvetted code.
- [COMMAND_EXECUTION]: The skill utilizes shell commands (
npx skills add) and executes a local Python script (scripts/request.py) to interface with an external skill marketplace. - [PROMPT_INJECTION]: The skill processes external data from the Alibaba Cloud skill marketplace (wuyingai.cn-shanghai.aliyuncs.com) to provide search results. This represents an indirect prompt injection surface where malicious search results could be used to trick an agent into recommending harmful repositories.
- Ingestion points: Search results returned as JSON from the
SearchPublicMarketSkillAPI action (SKILL.md). - Boundary markers: The skill includes instructions to the agent to avoid hallucinating names and to use only API-returned results, plus a requirement for explicit user confirmation before installation.
- Capability inventory: Shell command execution for installation (
npx skills add), Python script execution for searching. - Sanitization: The instructions do not specify validation or sanitization of the repository URLs or skill names retrieved from the external API.
Audit Metadata