The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions include an installation step using npx clawhub@latest install link-checker, which downloads code from a remote registry.
[REMOTE_CODE_EXECUTION]: The command npx clawhub@latest install link-checker executes code directly from an external source at runtime to install the necessary tooling.
[PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, making it vulnerable to indirect prompt injection.
Ingestion points: Data enters the agent context through web_fetch, screenshot, and general text analysis of external URLs provided by users or documents.
Boundary markers: There are no defined delimiters or instructions to the agent to disregard embedded commands within the fetched web content.
Capability inventory: The skill leverages web_fetch, screenshot, and image_edit (OCR) to retrieve and analyze remote content, which is then summarized or reported back to the agent.
Sanitization: While the skill performs security checks on the URLs themselves (e.g., domain reputation), it lacks mechanisms to sanitize or filter the textual content retrieved from those pages to prevent malicious instructions from being obeyed by the LLM.

link-checker