Modules§
- a2a
- adaptive
- adaptive_
mode_ policy - adaptive_
thresholds - agents
- archive
- attention_
model - Heuristic attention prediction model for LLM context optimization.
- bandit
- benchmark
- buddy
- budgets
- cache
- call_
graph - cli_
cache - codebook
- compression_
safety - compressor
- config
- consolidation_
engine - context_
deficit - context_
ledger - data_
dir - deep_
queries - Tree-sitter deep queries for extracting imports, call sites, and type definitions.
- deps
- editor_
registry - embedding_
index - Persistent, incremental embedding index.
- embeddings
- Embedding engine for semantic code search.
- entropy
- error
- events
- feedback
- filters
- gain
- gotcha_
tracker - graph_
context - Graph-driven context loading — automatically includes related files based on Property Graph proximity and token budgeting.
- graph_
index - graph_
provider - handoff_
ledger - heatmap
- hybrid_
search - Hybrid search combining BM25 (lexical) with dense vector search.
- import_
resolver - Import-to-file resolution (AST-driven import strings → project paths).
- index_
orchestrator - integrity
- Build-origin integrity verification and rebrand resistance.
- intent_
engine - intent_
protocol - knowledge
- knowledge_
embedding - Embedding-based Knowledge Retrieval for
ctx_knowledge. - language_
capabilities - limits
- litm
- llm_
feedback - loop_
detection - mcp_
manifest - memory_
lifecycle - Memory Lifecycle Management — consolidation, decay, compaction, archival.
- mode_
predictor - neural
- Neural context compression — trained models replacing heuristic filters.
- pathjail
- pathutil
- patterns
- pipeline
- pop_
pruning - portable_
binary - preservation
- project_
hash - property_
graph - Property Graph Engine — SQLite-backed code knowledge graph.
- prospective_
memory - protocol
- quality
- route_
extractor - safety_
needles - sandbox
- sanitize
- semantic_
cache - semantic_
chunks - Semantic Chunking with Attention Bridges.
- session
- setup_
report - signatures
- signatures_
ts - slow_
log - stats
- surprise
- Predictive Surprise Scoring — conditional entropy relative to LLM knowledge.
- symbol_
map - task_
briefing - task_
relevance - tdd_
schema - theme
- tokens
- updater
- vector_
index - version_
check - workflow
- wrapped