The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (CRITICAL): Confirmed malicious URL detection.
Evidence: The file references/index.md explicitly lists sql.md as a documentation resource.
Detection: Automated scanner URLite flagged sql.md as a blacklisted malicious URL (Alert ID: URE40E38848FEE8F9C-0200).
Analysis: The use of the .md Top-Level Domain (Moldova) for a domain name like sql.md appears to be a deliberate attempt to trick agents or users into interpreting an external network resource as a local markdown file.
[NO_CODE] (SAFE): No executable scripts, Python/Node.js packages, or other programmatic files were detected. The skill is limited to documentation, which reduces the immediate attack surface for binary exploitation but does not mitigate link-based risks.
[PROMPT_INJECTION] (LOW): The documentation structure presents an indirect prompt injection surface.
Ingestion points: references/index.md serves as an entry point for processing external documentation files.
Boundary markers: Absent; the skill provides no delimiters to prevent the agent from following instructions embedded in the linked files.
Capability inventory: None detected; no code capabilities like subprocess or network requests are provided by the skill itself.
Sanitization: None; the skill does not sanitize the contents of the referenced documentation.

postgresql