find-skills
Warn
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (MEDIUM): The skill installs external packages from GitHub or other sources using npx skills add. While vercel-labs/agent-skills is a trusted source, the skill allows installation from any repository.
- [COMMAND_EXECUTION] (MEDIUM): The instructions recommend using the -y flag (npx skills add -g -y), which bypasses confirmation prompts. This could lead to the silent installation of malicious tools if the agent is manipulated into selecting a harmful package.
- [Indirect Prompt Injection] (LOW): The skill has a surface for indirect injection via npx skills find. The agent processes search results (names/descriptions) from external sources to decide what to install. An attacker could poison repository metadata to trick the agent into installing a malicious skill. 1. Ingestion points: Command output from npx skills find. 2. Boundary markers: Absent. 3. Capability inventory: Subprocess calls via npx skills add. 4. Sanitization: None.
Audit Metadata