CodeAlmanac Architecture Audit
This folder captures the ongoing architecture audit requested on 2026-06-08. The audit is intentionally critical: existing code shape is treated as provisional until it earns its place.
Goal#
Determine the clean target architecture for the current CodeAlmanac codebase, including which current boundaries are correct, which are accidental special-case architecture, which features may not justify their cost, and which refactor slices should happen before future feature work.
Constraints#
- Do not modify source code during this investigation.
- Write findings frequently so the work survives context compaction.
- Use the repo wiki before inspecting related subsystems.
- Trust code over wiki if they disagree, but record disagreements.
- Use subagents or external research for independent critiques and prior-art checks where helpful.
Working Files#
worklog.md: running notes, decisions, open questions, and next steps.source-map.md: current repo map built from inspection.smells.md: suspected code smells and architectural debt.subagent-briefs.md: prompts sent to research/critique agents and returned summaries.target-architecture.md: emerging recommended shape and refactor roadmap.refactor-roadmap.md: proposed future implementation slices.research-notes.md: outside architecture research notes and takeaways.completion-audit.md: evidence that this read-only audit satisfied the request.reports/: full independent subagent reports.
Current Verdict#
The codebase does not need an enterprise-style folder rewrite. It does need a major cleanup of accumulated exceptions: sqlite-free CLI duplication, typed lifecycle start APIs, capture discovery consolidation, provider identity and runtime contract cleanup, shared query read models, viewer scope clarification, and a product decision on GitHub source access.