The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches documentation guidelines and templates from official Elastic sources, including elastic.co and the elastic/docs-content repository on GitHub. These are legitimate vendor resources used for content validation.
[COMMAND_EXECUTION]: The skill uses file system tools like Read, Grep, and Glob to access and analyze local documentation files as part of its core validation functionality.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from files or user-provided descriptions, creating a surface for indirect prompt injection attacks.
Ingestion points: Untrusted data enters the agent context through the $ARGUMENTS input and the Read tool, as described in SKILL.md.
Boundary markers: The instructions lack specific delimiters or warnings to the agent to disregard potential instructions embedded within the documentation being reviewed.
Capability inventory: The skill has access to tools including Read, Glob, WebFetch, and CallMcpTool (specifically for elastic-docs MCP), as defined in the SKILL.md frontmatter.
Sanitization: There are no explicit sanitization or filtering steps defined to process the content before it is analyzed by the agent.

docs-content-type-checker