Super Code Review: Full Project Quality Audit of Declared Approaches

[codenjoyme]

Overview

This issue is a request for a comprehensive, no-holds-barred quality audit of the entire repository. The goal is to verify that what the project declares it does is actually what it delivers — and to surface every gap, inconsistency, or missed opportunity for improvement.

The audit should be done by a coding agent scanning the entire repo. Be direct. Be brutal. Don't soften findings.

---

Scope of the Audit

1. Declared vs. Actual Module Structure

• Does every module under modules/ actually have both about.md and walkthrough.md?
• Are there modules that are empty folders, stubs, or skeleton files with no real content?
• Does each walkthrough.md actually teach what about.md claims it will teach?
• Are all modules referenced in training-plan.md actually present in modules/? Are there orphan modules not in the plan?

2. Instruction Files Quality

• Read every file in instructions/ and audit: are the instructions complete, actionable, and unambiguous?
• Do any instructions reference files, paths, or tools that don't exist in the repo?
• Are there circular or contradictory rules between instruction files?
• Does main.agent.md correctly describe all available instruction files? Are there files NOT listed there?

3. Consistency of Content Standards

• The project declares modules should be "5-7 minutes" and teach "one actionable skill" — verify this is true across all modules. Flag any modules that are too thin (no real content) or too fat (multi-skill dumps).
• Verify all module files are in English only (per the declared language rule in main.agent.md).
• Check that numbered lists use the declared auto-numbering format (1. 1. 1.) not sequential (1. 2. 3.).
• Check that all about.md files follow a consistent structure (title, description, topics, outcomes, prerequisites).

4. Training Plan Integrity

• Is training-plan.md up to date with the actual modules in modules/?
• Are module numbers sequential with no gaps or duplicates?
• Do module difficulty/complexity levels progress logically?
• Are prerequisites listed in about.md files actually pointing to existing modules?

5. Cross-Reference and Link Integrity

• Check all internal markdown links across the repo — do they resolve to existing files?
• Check all links in readme.md, quickstart.md, training-plan.md, motivation.md.
• Check cross-references between modules (e.g., "after completing module X") — do those modules exist?

6. Instructions Self-Consistency

• Does the project follow its own conventions? For example:
  • Does main.agent.md load correctly in every scenario it claims to handle?
  • Do *.agent.md files declare the keywords they handle? Are those keywords comprehensive?
  • Is the "FILE LOADING RULES (CRITICAL)" section in main.agent.md actually enforced in agent behavior patterns?

7. LND Output Quality

• Review lnd/output/ content against the rules in lnd/lnd-guide.md and lnd/rules/.
• Are generated LND files consistent with the source modules?
• Are images correctly organized and referenced?

8. Feedback Files

• Review all files in feedback/ — are the reported issues addressed or still open?
• Are there recurring patterns in feedback that suggest systemic problems?

9. General Improvement Proposals

For every finding above, don't just report — propose a concrete fix:

• If a file is missing: specify what it should contain.
• If a link is broken: propose the correct target.
• If content is inconsistent: propose the canonical standard.
• If an instruction is ambiguous: rewrite it.

---

Expected Output Format

Produce a structured report with sections matching the audit areas above. For each finding:

[FINDING] <short title>
Severity: critical | major | minor
Location: <file path or folder>
Problem: <clear description of the issue>
Proposed Fix: <concrete actionable recommendation>

At the end, produce a Summary Table with counts per severity and a Top 10 Priority Fixes list.

---

Notes for the Coding Agent

• Read every file. Don't sample. Don't skim.
• If a file seems fine on first glance — read it again against the declared standards.
• Cross-check main.agent.md against every file it references.
• Don't be polite about problems. A "looks mostly fine" is not useful. Find the cracks.
• No fix should be implemented in this issue — document only. PRs will be created separately per finding area.

[copilot-swe-agent]

The agent encountered an error and was unable to start working on this issue: Please try again later, or contact support if the problem persists. (Request id: 2224:20C191:9B5DDBD:73A944F:69E89FFE)