find-skills
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands via 'npx skills' to search for, install, and update agent capabilities. It recommends the use of the '-y' flag during installation, which bypasses interactive user confirmation prompts.
- [EXTERNAL_DOWNLOADS]: The skill facilitates downloading software packages from remote sources, including GitHub repositories and the skills.sh ecosystem.
- [REMOTE_CODE_EXECUTION]: By design, this skill installs and executes third-party code within the agent's environment via 'npx skills add'. It lacks built-in mechanisms for verifying the safety or origin of these external packages.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes user queries to search for external packages. Maliciously crafted package names or descriptions could be used to inject instructions into the agent's context.
- Ingestion points: User-provided search queries in SKILL.md processed by 'npx skills find'.
- Boundary markers: None; search queries are used directly to fetch external metadata.
- Capability inventory: Execution of 'npx skills' and installation of code from GitHub and other external sources.
- Sanitization: No input validation or output sanitization is performed on the search queries or the resulting package metadata.
Audit Metadata