The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to fetch documentation, configuration files (XML), and prompt templates from the NLWeb project's public GitHub repository (github.com/nlweb-ai/NLWeb). These resources are used for context and do not involve executable code downloads.
[COMMAND_EXECUTION]: Provides a curl command for developers to test custom tool routing on their local machine (localhost:8000). This is a standard practice for development and testing.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface.
Ingestion points: The framework's tool handlers (e.g., methods/your_tool.py) ingest and process user-provided query strings.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the provided implementation sketches.
Capability inventory: Handlers utilize context.retriever and context.ranker to interact with indexed data.
Sanitization: No specific input sanitization or validation logic is defined in the implementation guidance, as it focuses on framework architecture.

nlweb-tools-framework