io.github.alphaparkinc/genpark-context-window-token-pruner-skill (MCP Server)
This MCP server implements a priority-weighted context window token pruner with a FIFO eviction engine. It is described as protecting LLM prompt boundaries under prompt size limits, using token pruning and eviction strategies aligned with context management and prompt compression.
๐ ๏ธ Key Features
- Priority-weighted context window token pruning
- FIFO eviction engine
- Prompt-boundary protection for LLM prompt limits
๐ Use Cases
- Context management for LLM applications
- Preventing prompts from exceeding context window constraints
- Prompt compression through token pruning and eviction
โก Developer Benefits
- Deterministic FIFO eviction behavior
- Token selection driven by priority weighting
- LLM-focused approach to memory and prompt handling
โ ๏ธ Limitations
- No additional tool, interface, or configuration details were provided beyond the description and topics.