The Agent Skills Directory

[PROMPT_INJECTION]: The skill implements Retrieval-Augmented Generation (RAG) patterns that process external or user-provided data, creating a potential surface for indirect prompt injection.
Ingestion points: The scripts/generate_answer.py script retrieves context from namespaces, and scripts/upload_text.py uploads documents from JSON files.
Boundary markers: Absent. The scripts do not use specific delimiters or instructions to prevent the agent from following commands embedded within retrieved documents.
Capability inventory: The scripts/generate_answer.py script synthesizes natural language answers based on retrieved context, and scripts/search.py performs semantic searches.
Sanitization: Absent. The skill's scripts do not perform local filtering or sanitization of ingested content, relying instead on the backend Moorcheh API for safety.
[EXTERNAL_DOWNLOADS]: The skill requires external dependencies for its operation, which are vendor-provided or standard libraries.
Downloads the moorcheh-sdk Python package, which is the official software development kit for the platform.
References the requests library for API communication.
[COMMAND_EXECUTION]: The skill includes Python scripts designed to execute operations via the Moorcheh SDK, such as namespace management and data indexing.
These scripts (create_namespace.py, upload_text.py, etc.) are standard CLI tools for interacting with the service and do not contain arbitrary command execution vulnerabilities.

moorcheh