[Architecture] Thin CLI memory surface over shared memory application services

[orenlab]

[Architecture] Thin CLI memory surface over shared memory application services

Labels: enhancement, architecture, cli, memory
Priority: P2 (architecture triage)
Subject: codeclone/surfaces/cli/memory.py
Signal: responsibility_overload
Triage run: e28867ed (health 91/A)

Summary

codeclone/surfaces/cli/memory.py is a large CLI orchestration stack (argparse, observability, governance, trajectory, semantic, jobs) parallel to MCP’s _session_memory_mixin over the same codeclone.memory.* backends.

Environment

• Product/Service: CodeClone
• Version: 2.1.0a1 (run e28867ed)
• Area: CLI memory commands + Engineering Memory

Current State

1. Overloaded score 0.954, fan-out 29, unwind bottleneck candidate.
2. Impact blast: high (via workflow.py dispatch and extensive CLI memory test suite).
3. Module owns: memory_main, _dispatch_memory_command, trajectory/semantic/governance runners, store open helpers, observability wrapping.
4. MCP counterpart: codeclone/surfaces/mcp/_session_memory_mixin.py orchestrates the same memory backends via tools.

Expected Behavior

CLI memory commands should be thin adapters: parse argv → call shared memory application services → render via memory_render / trajectory/cli_render. Policy stays in codeclone.memory.governance, ide_governance, retrieval, ingest — not duplicated in the CLI surface.

Actual Behavior

One CLI module concentrates transport parsing and multi-lane command orchestration (records, trajectories, semantic rebuild, governance break-glass, jobs). New codeclone memory … subcommands expand the same overloaded surface rather than shared services.

Impact

Medium — No functional regression identified; risk is contract drift between CLI memory commands and MCP memory tools as both surfaces grow.

Proposed Direction

• Identify orchestration duplicated with _session_memory_mixin (search, init, trajectory status, semantic rebuild) and extract shared application functions under codeclone.memory or codeclone.memory.cli_service (name TBD).
• Keep CLI-specific concerns (argparse, console output, break-glass flags) in surfaces/cli/memory.py.
• Pattern reference: surfaces/cli/blast_radius.py delegates computation to codeclone.analysis.blast_radius.

Additional Context

• Architecture triage score: C+D → 25 (P2).
• Governance paths intentionally differ (CLI break-glass vs MCP IDE governance channel) — do not collapse those policies; reduce orchestration duplication only.
• Tests: tests/test_memory_cli*.py, tests/test_cli_memory_*.py.