WIP next tracks planing

conductor/tracks/deep_ast_context_pruning_20260306/spec.md

@@ -1,33 +1,28 @@
-# Track Specification: Deep AST-Driven Context Pruning (RAG for Code)
+# Track Specification: Deep AST-Driven Context Pruning
 
 ## Overview
-Implement deep ast-driven context pruning (rag for code) for Manual Slop application.
+Use tree_sitter to parse target file AST and inject condensed skeletons into worker prompts.
 
 ## Current State Audit
-### Already Implemented (DO NOT re-implement)
-- Existing backend functionality in src/ modules
-- Test coverage for core features
+### Already Implemented
+- file_cache.py with skeleton generation
+- outline_tool.py for code outlines
+- AST-based file analysis
 
-### Gaps to Fill (This Track Scope)
-This track addresses the gap between backend implementation and user-facing GUI/control.
+### Gaps to Fill
+- No automatic AST pruning based on edit target
+- Skeletons not customizable by user
 
 ## Goals
-- Implement deep ast-driven context pruning (rag for code)
-- Ensure test coverage
-- Follow existing code patterns
+- Parse target file AST before worker dispatch
+- Strip unrelated function bodies
+- Inject surgical skeleton to reduce tokens
 
 ## Functional Requirements
-- User-facing functionality as described in TASKS.md
-- Integration with existing backend
+- tree_sitter integration
+- Target function/method detection
+- Configurable skeleton depth
 
 ## Non-Functional Requirements
-- Performance: Maintain UI responsiveness
-- Tests: >80% coverage for new code
-
-## Architecture Reference
-- docs/guide_architecture.md
-- docs/guide_mma.md
-- docs/guide_tools.md
-
-## Out of Scope
-- Major refactoring of unrelated systems
+- <100ms parsing time per file
+- >50% token reduction target