boot-looplisted

# Boot Loop Bootstrap or overhaul a project's agent system from empirical evidence gathered by parallel investigative workers. `boot-loop` probes the codebase through focused investigation tasks, collects the friction notes they produce, and runs one `solid-loop` pass on them. ## Workflow ### 1. Scope Investigative Tracer Jobs Select 2 to 4 distinct, high-impact investigation targets representing primary areas of the repository: - **Build & Test Reality**: Run test suites, typecheckers, or linters to identify local execution quirks and missing setup steps. - **Architectural Seams**: Trace a core user flow or data pipeline to identify primary entry points and boundaries. - **Conventions & Idioms**: Inspect typical module structure, configuration conventions, and dependency patterns. Each job must have a concrete, falsifiable objective. ### 2. Dispatch Independent Workers (`aaaav-do`) Dispatch isolated subagents or background workers for each target: - Each worker follows **`aaaav-do`** (Align → Advance → Anchor → Act → Verify). - Each worker records and classifies friction notes as `aaaav-do` defines them, then returns the classified notes instead of invoking `solid-loop`. Wait for workers to complete their designated investigation runs. ### 3. Merge Friction Notes Collect the classified notes from all completed workers and merge notes whose `Found` lines state the same fact. ### 4. Run One `solid-loop` Pass Apply the merged notes through a single `solid-loop` p