The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates the ingestion of data from external, untrusted URLs (e.g., in metadata.md and page-text.md), which creates a surface for indirect prompt injection where malicious instructions could be embedded in the source content. \n
Ingestion points: mql function calls in SKILL.md and workflow files like page-markdown.md. \n
Boundary markers: Not present; the skill does not instruct the agent to use delimiters or specific warnings to ignore instructions in fetched content. \n
Capability inventory: The skill enables network requests and structured data extraction via the Microlink API. \n
Sanitization: Absent; content from target URLs is passed to the agent without filtering or sanitization. \n- [REMOTE_CODE_EXECUTION]: Documents the use of the function and evaluate parameters of the Microlink API, which allow executing JavaScript within a remote headless browser environment. This is an intended feature of the well-known Microlink service. \n- [EXTERNAL_DOWNLOADS]: Recommends the installation of the @microlink/mql and @microlink/cli packages from the official NPM registry and references the esm.sh CDN.

microlink-api