The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses user-provided strings directly in shell commands, creating a potential command injection surface.
Ingestion points: User-supplied search terms are captured in the [query] placeholder within the find-skills logic (SKILL.md).
Boundary markers: No delimiters or instructions to treat the user query as data rather than code are provided to the agent.
Capability inventory: The skill utilizes the npx shell command to execute skills find and skills add via the system terminal.
Sanitization: There is no evidence of input validation, escaping, or sanitization before the query is passed to the shell.
[REMOTE_CODE_EXECUTION]: The skill instructs the agent to install third-party code packages from the internet.
The instructions explicitly recommend using the -y flag (npx skills add <package> -g -y), which bypasses the user confirmation prompt that normally serves as a security checkpoint for code installation.
[COMMAND_EXECUTION]: The skill relies on shell command execution as its primary mechanism.
It uses npx to run the skills utility for searching, adding, and updating packages on the local system.
[EXTERNAL_DOWNLOADS]: The skill interacts with external repositories and services.
It fetches skill information and installation payloads from https://skills.sh/ and references GitHub repositories.
References to vercel-labs/agent-skills target a known organization's repository for best practices.

find-skills