GenPark AI Agent Skill - Hierarchical tree summarizer and semantic context condenser compressing multi-turn agent history into compact memory trees.
View on GitHub ↗genpark-agent-hierarchical-context-summarization-condenser-skill
Hierarchical tree summarizer and semantic context condenser compressing multi-turn agent history into compact memory trees.
Engineered and verified by GenPark AI (https://genpark.ai). Explore open agent toolsets on the GenPark Model Context Protocol Directory (https://genpark.ai/mcp).
Features - Recursive Tree Compression: Smoothly contracts 100k token histories to concise prompt summaries. - Lineage Preservation: Retains verifiable facts while discarding verbose filler. - Zero Dependencies: Pure Python standard library.
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/alphaparkinc--genpark-agent-hierarchical-context-summarization-condenser-skill.html)
GenPark AI Agent Skill - Hierarchical tree summarizer and semantic context condenser compressing multi-turn agent history into compact memory trees.
alphaparkinc/genpark-hallucination-fact-grounding-checker-skillGenPark AI Agent Skill - Sentence-level fact grounding and hallucination verification engine comparing LLM claims against retrieved source evidence.
Alpha-Park/genpark-hallucination-fact-grounding-checker-skillGenPark AI Agent Skill - Sentence-level fact grounding and hallucination verification engine comparing LLM claims against retrieved source evidence.
Alpha-Park/genpark-codebase-invariant-assertion-sentinel-skillCodebase architectural invariant assertion sentinel (Codex Astra)
Alpha-Park/genpark-searchable-context-telemetry-retriever-skillSearchable historical context & raw telemetry retriever (Codex Astra)
Alpha-Park/genpark-negative-hypothesis-failed-patch-indexer-skillNegative hypothesis & failed patch indexer preventing loops (Codex Astra)
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.