code-quality-scan

Files (1)

• SKILL.md

SKILL.md (first lines)

# Code Quality Scan

This skill defines the Sentinel scanning workflow — a systematic, cross-repo quality audit that discovers lint errors, dead code, duplication, config inconsistencies, and dependency drift across all AprovanLabs repositories, then creates scoped improvement issues.

---

## Per-Repo Scan Config

| Repo | Language | Scan Tools | Notes |
|---|---|---|---|
| core | TypeScript | eslint, tsc, knip, jscpd | Monorepo with pnpm |
| registry | TypeScript | eslint, tsc, knip, jscpd | Monorepo with pnpm |
| apprentice | TBD | TBD | Investigate stack before scanning |
| patchwork | TBD | TBD | Investigate stack before scanning |
| zolvery | TBD | TBD | Investigate stack before scanning |
| projects | TBD | TBD | Investigate stack before scanning |
| aprovan.com | TypeScript | eslint, build | Static website |
| data-science | Python | flake8, mypy, jscpd | Python stack |

---

## Phase 1 — Repo Discovery

1. List all AprovanLabs repos from the workspace:

```bash
multica repo checkout https://github.com/AprovanLabs/core
multica repo checkout https://github.com/AprovanLabs/registry
multica repo checkout https://github.com/AprovanLabs/apprentice
multica repo checkout https://github.com/AprovanLabs/patchwork
multica repo checkout https://github.com/AprovanLabs/zolvery
multica repo checkout https://github.com/AprovanLabs/projects
multica repo checkout https://github.com/AprovanLabs/aprovan.com
multica repo checkout https://github.com/AprovanLabs/data-science
```

2. For each repo, identify its type:
   - **TypeScript**: has `tsconfig.json` and/or `package.json` with TypeScript deps
   - **Python**: has `pyproject.toml`, `setup.py`, or `.python-version`
   - **Mixed**: both TypeScript and Python indicators
3. Select applicable scan tools from the per-repo config table above
4. For repos marked TBD: run a quick investigation (`ls`, read `package.json`/`pyproject.toml`), update the config table, and note findings

---

## Phase 2 — Per-Repo Scanning

For each repo, run the applicable scan tools. Capture all findings with file paths, line numbers, and severity.

### 2.1 Deslopify Scan

If `deslopify` is available and configured for the repo:

```bash
deslopify scan --path .
deslopify backlog
```

Capture the backlog items as findings.

### 2.2 Lint (TypeScript repos)

```bash
pnpm lint
```

Capture every lint error with file, line, rule name, and message. Warnings are informational; errors are high-severity.

### 2.3 Type Check (TypeScript repos)

```bash
pnpm typecheck
```

Capture every type error with file, line, and message. Type errors are high-severity.

### 2.4 Dead Code Detection

```bash
npx knip --no-progress
```

Capture unused exports, unreachable files, and duplicate dependencies. Severity: medium.

### 2.5 Duplicate Code Detection

```bash
npx jscpd --threshold 5
```

Capture duplicate blocks with file pairs and line ranges. Severity: medium.

For Python repos, use AST-based duplication tools or manual review.

### 2.6 Python Lint and Type Check

```bash
flake8 . --count --statistics
mypy . --ignore-missing-imports
```

### 2.7 Missing Shared Config Check

Check whether the repo uses shared AprovanLabs config packages:

- `@aprovan/eslint-config` — shared ESLint config
- `@aprovan/tsconfig` — shared TypeScript config
- `@aprovan/prettier-config` — shared Prettier config

```bash
cat package.json | python3 -c "
import json, sys
pkg = json.load(sys.stdin)
deps = {**pkg.get('dependencies', {}), **pkg.get('devDependencies', {})}
shared = ['@aprovan/eslint-config', '@aprovan/tsconfig', '@aprovan/prettier-config']
for s in shared:
    if s not in deps:
        print(f'MISSING: {s}')
    else:

From AprovanLabs/registry, MIT.

More from AprovanLabs/registry