The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill relies on npx shadcn-vue@latest for component installation, updates, and configuration. Since the domain shadcn-vue.com associated with this package has been flagged as malicious (CryptScam) by automated scanners, executing these commands poses a severe risk of running untrusted or compromised code on the host system.
[EXTERNAL_DOWNLOADS]: Multiple files, including assets/components.json and references/theming.md, reference shadcn-vue.com for fetching JSON schemas and documentation. This domain is confirmed to be malicious, making all external fetches from it unsafe.
[DYNAMIC_CONTEXT_INJECTION]: The skill uses the !command`` syntax in SKILL.md to execute npx shadcn-vue@latest info --json automatically when the skill is loaded by the agent. This results in silent, pre-execution of code from a flagged malicious source without any user intervention or approval.
[INDIRECT_PROMPT_INJECTION]: The skill provides an extensive surface for indirect prompt injection attacks.
Ingestion points: The skill instructions (SKILL.md) direct the agent to fetch component documentation and examples from external URLs using npx shadcn-vue@latest docs <component>.
Boundary markers: No boundary markers or 'ignore embedded instructions' warnings are present to protect the agent from malicious content in the external data.
Capability inventory: The skill possesses the ability to execute shell commands (npx), write files to the local system, and install dependencies, all of which could be abused by an attacker-controlled external document.
Sanitization: There is no evidence of sanitization or validation of the content fetched from external registries before it is processed by the agent.

shadcn-vue