mcpField
L
ln-645-architecture-modernization-auditor

ln-645-architecture-modernization-auditor

github.com/levnikolaevichTesting & quality58470 installsUpdated 2d ago
Repository

About ln-645-architecture-modernization-auditor

A code-audit skill for finding architecture-level simplification and modernization opportunities. It is intended for coding agents reviewing repositories and produces evidence-based, migration-aware findings rather than changing code.

Key features

  1. 01Detect obsolete custom mechanisms
  2. 02Review overbuilt extension points
  3. 03Find parallel architecture mechanisms
  4. 04Verify candidates in context
  5. 05Score and write audit findings

Use cases

  1. 01Review a codebase for removable architecture indirection
  2. 02Assess a single-use plugin or registry design
  3. 03Find overlapping dispatch or service composition patterns

MCP support

The skill prefers hex-line for local code reads and hex-graph for relationship analysis when available. It falls back to built-in repository tools if MCP is unavailable or unsupported.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-645-architecture-modernization-auditor

After installing, type/ln-645-architecture-modernization-auditorin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

references/
audit_scoring.md
audit_summary_contract.md
audit_worker_core_contract.md
mcp_integration_patterns.md
mcp_tool_preferences.md
two_layer_detection.md
references/templates/
audit_worker_report_template.md

Details

From the project README

Paths: File paths (references/, ../ln-*) are relative to this skill directory.

Architecture Modernization Auditor

Type: L3 Worker

L3 Worker that finds architecture-level simplification and modernization opportunities.

Purpose & Scope

  • Audit obsolete custom architectural mechanisms, overbuilt extension points, and avoidable architecture indirection
  • Identify mechanisms that increase maintenance cost without current architectural value
  • Generate migration-safe modernization findings
  • Emit SIMPLIFY_ARCHITECTURE, RETIRE_CUSTOM_MECHANISM, or CONSOLIDATE_EXTENSION_POINT
  • Calculate compliance score (X/10)

Inputs

MANDATORY READ: Load references/audit_worker_core_contract.md and references/mcp_tool_preferences.md. Tool policy: You may run as an isolated subagent where host AGENTS.md is not in scope, so default to hex-line MCP first for file reads, search, and edits. Load references/mcp_integration_patterns.md only when MCP behavior is unclear.

Receives contextStore with: tech_stack, architecture, codebase_root, output_dir, domain_mode, scan_path.

Use hex-graph first when references, architecture metrics, or implementation relationships materially improve modernization findings. Use hex-line first for local code reads when available. If MCP is unavailable, unsupported, or not indexed, continue with built-in Read/Grep/Glob/Bash and state the fallback in the report.

Workflow

Detection policy: use two-layer detection (candidate scan, then context verification); load references/two_layer_detection.md only when the verification method is ambiguous.

  1. Parse context -- determine scan_root from domain-aware scope or codebase root
  2. Find modernization candidates (Layer 1)
    • Custom plugin/extension systems with one implementation
    • Hand-built service locators, registries, dispatchers, buses, adapters, or factories with limited use
    • Legacy architecture mechanisms with deprecation comments or unused compatibility branches
    • Parallel architectural mechanisms solving the same problem
  3. Verify architectural value (Layer 2)
    • Is the mechanism required by the framework or deployment model? -> skip
    • Is it a public extension API with external consumers? -> downgrade or skip
    • Does it isolate a volatile boundary with multiple real implementations? -> skip
    • Is the replacement only a package swap or utility rewrite? -> skip
  4. Collect findings with severity, location, action, effort, migration risk, and recommendation
  5. Calculate score using references/audit_scoring.md
  6. Write report to {output_dir}/ln-645--{identifier}.md
  7. Return summary per references/audit_summary_contract.md

Audit Rules

1. Obsolete Custom Mechanisms

What: Custom architecture mechanisms kept after the project no longer needs them

Detection:

  • Search for Legacy, Compat, Shim, Adapter, Registry, Plugin, Extension, ServiceLocator, Dispatcher, EventBus
  • Check references and implementation count
  • Read deprecation/TODO comments and ADRs when present

Severity: HIGH for critical-path mechanisms, MEDIUM for broad but non-critical mechanisms, LOW for isolated cleanup

Action: RETIRE_CUSTOM_MECHANISM

2. Overbuilt Extension Points

What: Extension or plugin architecture with one implementation and no active variation point

Detection:

  • Interfaces/abstract classes used only by one implementation
  • Factories or registries with one registered type
  • Configuration-driven plugin loading with one hardcoded plugin

Severity: MEDIUM by default, HIGH when the indirection hides critical flow

Action: CONSOLIDATE_EXTENSION_POINT

3. Parallel Architectural Mechanisms

What: Two or more architecture mechanisms solve the same problem in different modules

Detection:

  • Multiple dispatch/registry/event mechanisms
  • Multiple service composition styles
  • Multiple adapter conventions for the same boundary

Severity: HIGH when it creates inconsistent behavior, MEDIUM when it mainly increases maintenance cost

Action: SIMPLIFY_ARCHITECTURE

Scoring Algorithm

MANDATORY READ: Load references/audit_scoring.md.

Output Format

MANDATORY READ: Load references/templates/audit_worker_report_template.md.

Write JSON summary per references/audit_summary_contract.md. In managed mode the caller passes both runId and summaryArtifactPath; in standalone mode the worker generates its own run-scoped artifact path per shared contract.

Write report to {output_dir}/ln-645--{identifier}.md with category: "Architecture Modernization" and checks: obsolete_mechanisms, overbuilt_extension_points, parallel_mechanisms.

When summaryArtifactPath is absent, write the standalone runtime summary under .hex-skills/runtime-artifacts/runs/{run_id}/evaluation-worker/{worker}--{identifier}.json and optionally echo the same summary in structured output.

Critical Rules

Apply the already-loaded references/audit_worker_core_contract.md.

  • Do not auto-fix: Report only, never rewrite architecture
  • Unique angle: Audit architecture-level modernization only. Do not audit package health, CVEs, generic utility replacement, code complexity, or runtime configuration hygiene.
  • No package recommendations: Do not search for OSS replacements or recommend dependency swaps.
  • Evidence required: Every modernization finding must include implementation count, reference evidence, and migration risk.
  • Action required: Every finding uses SIMPLIFY_ARCHITECTURE, RETIRE_CUSTOM_MECHANISM, or CONSOLIDATE_EXTENSION_POINT.

Definition of Done

Apply the already-loaded references/audit_worker_core_contract.md.

  • contextStore parsed successfully
  • scan_root determined
  • obsolete custom mechanisms checked
  • overbuilt extension points checked
  • parallel architectural mechanisms checked
  • Layer 2 architectural value verification applied
  • Findings collected with severity, location, action, effort, migration risk, and recommendation
  • Score calculated per references/audit_scoring.md
  • Report written to {output_dir}/ln-645--{identifier}.md (atomic single Write call)
  • Summary written per contract

Version: 1.0.0 Last Updated: 2026-02-26

Comments

Markdown supported