The Agent Skills Directory

[SAFE]: The skill uses specific tools provided by the Kleap MCP server to read, write, and publish website source files. These operations are essential for the stated purpose of website editing and do not involve unauthorized data access or exfiltration.- [SAFE]: The skill exhibits an indirect prompt injection surface because it reads external file content (Astro components and JSON data) which could theoretically contain instructions intended to influence the agent. This is documented as a safe observation for this tool type. * Ingestion points: File contents are retrieved using the read_files tool (SKILL.md). * Boundary markers: No specific delimiters or safety warnings are defined for the ingested content. * Capability inventory: The agent has access to write_files, modify_app, and publish_app for modifying and deploying site content. * Sanitization: No content validation or sanitization steps are specified in the instructions.

kleap-edit