The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data from mydlink.com subdomains which could contain untrusted instructions. * Ingestion points: Tools like la_mydlink_com__read_page and generic unbrowse.scrape tools described in SKILL.md fetch content from external URLs. * Boundary markers: The skill does not define specific delimiters or instructions for the agent to ignore content within the fetched data. * Capability inventory: The skill provides capabilities for network interaction and browser-like actions via unbrowse.run and unbrowse.browse.act. * Sanitization: There is no evidence of automated content sanitization; however, the skill explicitly instructs the agent to get user confirmation for sensitive operations like posting, buying, or deleting.
[EXTERNAL_DOWNLOADS]: The skill configuration and installation instructions point to resources hosted on the vendor's domain, unbrowse.ai. * Installation instructions in README.md and SKILL.md suggest using claude mcp add and npx skills add with URLs pointing to https://unbrowse.ai. * These references are consistent with the documented author's infrastructure and represent standard functionality.

mydlink