The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill uses the command npx skills init <skill-name>, which downloads and executes a package from the npm registry. Using unversioned or generic package names from public registries poses a risk of executing malicious code if the package name is hijacked or untrusted.
[EXTERNAL_DOWNLOADS]: The workflow involves fetching external code from the npm registry during the scaffolding step.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands for scaffolding and for running local verification scripts ('skill check') within the repository.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it transforms user-provided prompts and ideas into procedural instructions for new skills. * Ingestion points: User-provided goal, idea, or workflow (SKILL.md). * Boundary markers: Absent; there are no instructions to delimit or ignore instructions within the user's input. * Capability inventory: The agent can write files, create directories, and execute shell commands. * Sanitization: Absent; the skill does not specify any validation or filtering of user input before it is incorporated into the generated skill body.

create-skill