entroly_mcp
DNA.expression(A specialized context and memory optimization suite. Entroly (v1.0.0) provides 10 tools for mathematically optimal context selection, semantic recall, and session checkpointing. Built for Minds requiring high-fidelity long-term memory and efficient token usage.)
Serialize and save the current session state (checkpoints) for persistent cross-session memory.
Provide a per-fragment scoring breakdown to explain why specific code or documentation was included or excluded.
Retrieve real-time session statistics, including token usage, cost savings, and coverage percentages.
Run static analysis to identify code clones, dead symbols, 'god files', and other codebase health metrics.
Select the mathematically optimal subset of context that fits within a specified token budget.
Predict and pre-load context likely to be needed based on current query intent and dependency graphs.
Perform sub-linear semantic recall across indexed fragments using multi-probe Locality Sensitive Hashing (LSH).
Record the outcome of a context selection to feed the reinforcement learning loop (PRISM) and improve future results.
Store context fragments with auto-deduplication, entropy scoring, and dependency linking.
Restore a session from a previously saved checkpoint or state file.