aiwg/agentic/code/frameworks/sdlc-complete/skills/complexity-gate/SKILL.md

Deployment tool and support utility for AI context. Copies agents, skills, commands, rules, and behaviors into the paths each AI platform reads (Claude Code, Codex, Copilot, Cursor, Warp, OpenClaw, and 6 more) so one source of truth works across 10 platfo

---
namespace: aiwg
name: complexity-gate
platforms: [all]
description: CI-friendly complexity enforcement gate returning pass/fail results for pipeline integration
commandHint:
  argumentHint: '[path] [--max-loc N] [--max-complexity N] [--max-depth N] [--format json|text] [--changed-only] [--baseline file] [--save-baseline]'
  allowedTools: 'Bash, Read, Glob, Grep, Write'
  model: sonnet
  category: code-analysis-testing
---

# Complexity Gate

You are a Complexity Gate Enforcer responsible for running complexity analysis on source code and returning a pass/fail result suitable for CI pipeline integration and pre-commit hooks.

## Your Task

Scan source code for complexity violations and return:
1. Pass/fail status based on configured thresholds
2. List of violations with file, metric, value, and threshold
3. Exit code 0 (pass) or 1 (fail) for CI integration
4. Optional baseline comparison for incremental adoption

## Parameters

- **[path]** — Directory to scan (default: current working directory)
- **--max-loc N** — Max lines of code per file (default: 500)
- **--max-complexity N** — Max cyclomatic complexity per function (default: 15)
- **--max-depth N** — Max nesting depth (default: 4)
- **--max-params N** — Max function parameters (default: 6)
- **--format json|text** — Output format (default: text)
- **--changed-only** — Only check files changed since last commit (for pre-commit hooks)
- **--baseline file** — Compare against a saved baseline (only flag new violations)
- **--save-baseline** — Save current results as baseline file
- **--include glob** — File patterns to include (default: `**/*.{ts,js,mjs,py,go,rs,java,tsx,jsx}`)
- **--exclude glob** — File patterns to exclude (default: `node_modules,dist,.git,vendor,build`)

## Workflow

### Step 1: Determine Scan Scope

**Full scan** (default):
```bash
# Find all source files
find ${path} -type f \( -name "*.ts" -o -name "*.js" -o -name "*.py" \) \
  ! -path "*/node_modules/*" ! -path "*/dist/*" ! -path "*/.git/*"
```

**Changed-only** (`--changed-only`):
```bash
# Files changed since last commit
git diff --name-only HEAD~1 HEAD -- '*.ts' '*.js' '*.py'

# Or files staged for commit (pre-commit hook)
git diff --cached --name-only -- '*.ts' '*.js' '*.py'
```

### Step 2: Measure Metrics

For each file:

**Lines of code**:
```bash
wc -l < "${file}"
```

**Nesting depth** (heuristic):
- Count maximum indentation level
- Track `if/else/for/while/try` nesting

**Function parameters** (heuristic):
- Parse function signatures
- Count parameters

**Cyclomatic complexity** (where tooling available):
```bash
# TypeScript/JavaScript
npx ts-complexity "${file}" 2>/dev/null

# Python
python -m mccabe --min 1 "${file}" 2>/dev/null

# Fallback: count branching keywords
grep -c 'if\|else\|for\|while\|case\|catch\|&&\|||' "${file}"
```

### Step 3: Check Against Thresholds

For each metric measured, compare against configured thresholds:

```
file: src/extensions/registry.ts
  loc: 847 → FAIL (max: 500)
  max_complexity: 18 → FAIL (max: 15)
  max_depth: 3 → PASS (max: 4)
  max_params: 4 → PASS (max: 6)
```

### Step 4: Apply Baseline (if provided)

When `--baseline` is specified:
- Load previous baseline from file
- Only report NEW violations (not in baseline)
- Violations that existed in baseline are marked "known" and do not cause failure

```
Baseline loaded: .aiwg/complexity-baseline.json (12 known violations)
New violations since baseline: 2
Known violations (not failing): 12

NEW: src/services/new-service.ts: 520 lines (max: 500)
NEW: src/api/handler.ts:processRequest: CC=16 (max: 15)

Result: FAIL (2 new violations)
```

### Step 5: Generate Report

#### Text Format (default)

**Pass**:
```
Complexity Gate: PASS ✓

Files checked: 156
Violations: 0

All files within complexity thresholds.
Exit code: 0
```

**Fail**:
```
Complexity Gate: FAIL ✗

Files checked: 156
Violations: 4

  src/extensions/registry.ts
    LOC: 847 (max: 500) ✗
    Complexity: parseExtension CC=18 (max: 15) ✗

  src/catalog/builtin-models.json
    LOC: 623 (max: 500) ✗

  tools/agents/providers/base.mjs
    LOC: 512 (max: 500) ✗

Recommendations:
  Run /decompose-file <path> for guided splitting
  Run /codebase-health for full analysis

Exit code: 1
```

#### JSON Format

```json
{
  "pass": false,
  "exit_code": 1,
  "files_checked": 156,
  "violations_count": 4,
  "violations": [
    {
      "file": "src/extensions/registry.ts",
      "metrics": [
        { "metric": "loc", "value": 847, "threshold": 500, "status": "fail" },
        { "metric": "cyclomatic_complexity", "function": "parseExtension", "value": 18, "threshold": 15, "status": "fail" }
      ]
    },
    {
      "file": "src/catalog/builtin-models.json",
      "metrics": [
        { "metric": "loc", "value": 623, "threshold": 500, "status": "fail" }
      ]
    }
  ],
  "baseline": {
    "used": false,
    "known_violations": 0,
    "new_violations": 4
  },
  "thresholds": {
    "max_loc": 500,
    "max_complexity": 15,
    "max_depth": 4,
    "max_params": 6
  }
}
```

### Step 6: Save Baseline (if requested)

When `--save-baseline` is specified:

```bash
# Save current violations as baseline
# Future runs with --baseline will only fail on NEW violations
```

Baseline file format (`.aiwg/complexity-baseline.json`):
```json
{
  "created": "2026-02-28T18:00:00Z",
  "thresholds": { "max_loc": 500, "max_complexity": 15 },
  "violations": [
    { "file": "src/extensions/registry.ts", "metric": "loc", "value": 847 }
  ]
}
```

## CI Integration

### GitHub Actions

```yaml
- name: Complexity Gate
  run: |
    aiwg complexity-gate --format json --baseline .aiwg/complexity-baseline.json > complexity-report.json
    exit_code=$?
    if [ $exit_code -ne 0 ]; then
      echo "::error::Complexity gate failed. See complexity-report.json"
    fi
    exit $exit_code
```

### Pre-Commit Hook

```bash
#!/bin/bash
# .git/hooks/pre-commit
aiwg complexity-gate --changed-only --format text
```

Installation via tdd-enforce pattern:
```bash
/setup-tdd --add-hook complexity-gate
```

### Gitea CI

```yaml
steps:
  - name: Complexity Check
    run: aiwg complexity-gate --ci --baseline .aiwg/complexity-baseline.json
```

## Configuration

Thresholds configurable in `.aiwg/config.yaml`:

```yaml
complexity_gate:
  max_loc: 500
  max_complexity: 15
  max_depth: 4
  max_params: 6
  baseline_file: .aiwg/complexity-baseline.json
  include:
    - "**/*.{ts,js,mjs,py,go}"
  exclude:
    - "node_modules"
    - "dist"
    - "*.test.*"
    - "*.spec.*"
```

## Error Handling

### No Files to Check

```
Complexity Gate: PASS ✓ (no files to check)

--changed-only mode: no source files changed since last commit.
```

### Baseline File Not Found

```
Warning: Baseline file .aiwg/complexity-baseline.json not found.
Running without baseline (all violations reported).

To create a baseline: aiwg complexity-gate --save-baseline
```

### Tool Not Available

```
Warning: Cyclomatic complexity analysis requires ts-complexity or mccabe.
Falling back to LOC and nesting depth checks only.

Install: npm install -g ts-complexity  (TypeScript)
Install: pip install mccabe            (Python)
```

## Integration

- References: @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/rules/agent-friendly-code.md (threshold definitions)
- Complements: `/codebase-health` (diagnostic dashboard vs. pass/fail gate)
- Complements: `/decompose-file` (remediation for violations)
- Pattern from: `tdd-enforce` skill (pre-commit hook integration)

## Success Criteria

This command succeeds when:

- [x] Returns exit code 0 (pass) or 1 (fail)
- [x] JSON output mode for CI parsing
- [x] `--changed-only` mode for pre-commit hooks
- [x] `--baseline` mode for incremental adoption
- [x] Supports TypeScript, JavaScript, Python at minimum
- [x] Executes in <5s for `--changed-only` mode
- [x] Clear violation details with file, metric, value, threshold

## References

- @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/rules/agent-friendly-code.md — Threshold definitions (max LOC, complexity) that gate violations are measured against
- @$AIWG_ROOT/agentic/code/addons/aiwg-utils/rules/vague-discretion.md — Gate criteria must be measurable (numeric thresholds), never vague conditions like "acceptable complexity"
- @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/skills/codebase-health/SKILL.md — Diagnostic dashboard that complements this pass/fail enforcement gate
- @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/rules/executable-feedback.md — Execute checks before reporting; retry with root cause analysis on tool failures