/code:completeness

Multi-agent analysis for forgotten features, dead handlers, stub code, and incomplete implementations

/code:completeness

Deploy multiple specialized completeness analyzers in parallel to find forgotten features, dead handlers, stub code, and incomplete implementations, then synthesize results through consensus voting into a prioritized Completeness Audit Report.

Quick Start

/agileflow:code:completeness app/
/agileflow:code:completeness . DEPTH=deep
/agileflow:code:completeness src/ FOCUS=handlers,routes

Arguments

Argument	Values	Default	Description
`[file\|directory]`	file or directory	`.`	What to analyze
`DEPTH`	quick, deep, ultradeep	quick	Analysis depth: quick = core 5 analyzers, deep = all 7 analyzers, ultradeep = separate tmux sessions
`FOCUS`	handlers, routes, api, stubs, state, imports, conditional, all	all	Which analyzers to deploy (comma-separated)
`MODEL`	haiku, sonnet, opus	haiku	Model for analyzer subagents

How It Works

Parse arguments - Determine target, depth, and focus areas
Deploy analyzers in parallel - 5-7 specialized agents depending on depth
Collect all findings - Wait for all analyzers to complete
Run consensus coordinator - Validate findings and prioritize by severity
Generate Completeness Audit Report - Actionable recommendations saved to docs/08-project/completeness-audits/

Core Analyzers (Quick Mode - 5)

Analyzer	Finds
Handlers	Dead/empty event handlers, console-only handlers
Routes	Dead navigation, broken links, missing pages
API	Frontend-backend endpoint mismatches, orphaned endpoints
Stubs	TODO/FIXME comments, empty function bodies, mock data in production
State	Unused useState/useReducer, orphaned context providers

Deep Mode Analyzers (2 Additional)

Analyzer	Finds
Imports	Dead exports, unused dependencies, orphaned modules
Conditional	Dead feature flags, unreachable code, commented-out features

Severity Scale

Level	Definition	Example
BROKEN	Visibly broken in production - user encounters crash or non-functional element	Empty onClick handler, API call to non-existent endpoint
INCOMPLETE	Feature exists but silently does nothing or loses data	Form submits but handler is noop, state set but never read
PLACEHOLDER	Development stub shipped to production	`TODO: implement`, `throw new Error('Not implemented')`, hardcoded mock data
DORMANT	Unused code that should be alive or removed	Dead export, hardcoded false feature flag, commented-out feature

Confidence Levels

Level	Definition	Priority
CONFIRMED	2+ analyzers agree on finding	High
LIKELY	1 analyzer with strong evidence	Medium
INVESTIGATE	1 analyzer with weak evidence	Low

Examples

# Quick audit of app directory (core 5 analyzers)
/agileflow:code:completeness app/
 
# Deep audit with all 7 analyzers
/agileflow:code:completeness . DEPTH=deep
 
# Focus on specific incomplete features
/agileflow:code:completeness src/ FOCUS=handlers,routes
 
# Check for stubs and unused state only
/agileflow:code:completeness components/ FOCUS=stubs,state
 
# Comprehensive full audit with deep model
/agileflow:code:completeness . DEPTH=deep FOCUS=all MODEL=sonnet

Example Output

🔍 Completeness Audit: app/
══════════════════════════════════════════════════════════════

Deploying 5 completeness analyzers (quick mode)...
✓ Handlers Analyzer
✓ Routes Analyzer
✓ API Analyzer
✓ Stubs Analyzer
✓ State Analyzer

Running consensus...
✓ Consensus complete
✓ Project type detected: Full-stack Web Application

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📊 COMPLETENESS SUMMARY
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

| Severity    | Count | User Impact            |
|-------------|-------|------------------------|
| Broken      | 2     | user-blocking          |
| Incomplete  | 3     | user-confusing         |
| Placeholder | 1     | data-silent            |
| Dormant     | 4     | developer-only         |

Total: 10 findings (3 intentional exclusions)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🚨 SHIP BLOCKERS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

1. Empty onClick handler on "Delete Account" button [CONFIRMED]
   Location: components/Settings.tsx:45
   Severity: BROKEN | Impact: user-blocking
   Remediation: Implement delete API call or hide button

2. fetch('/api/payments') but route doesn't exist [CONFIRMED]
   Location: pages/checkout.tsx:28
   Severity: BROKEN | Impact: user-blocking
   Remediation: Create API route or disable checkout

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⚠️  FIX BEFORE RELEASE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

3. Form onSubmit only sets loading state, never calls API [LIKELY]
   Location: components/ContactForm.tsx:22
   Severity: INCOMPLETE | Impact: user-confusing

4. useState for searchResults setter called but value never rendered [LIKELY]
   Location: pages/search.tsx:8
   Severity: INCOMPLETE | Impact: data-silent

[Full report saved to docs/08-project/completeness-audits/]

Quick Depth Comparison

Aspect	quick	deep	ultradeep
Analyzers	5 (core)	7 (all)	7 (all, parallel tmux)
Execution	Sequential	Sequential	Parallel sessions
Severity Focus	BROKEN/INCOMPLETE	All 4 levels	All 4 levels
Time	<5min	5-10min	5-10min (parallel)
Cost	Low	Medium	Medium (parallelized)

What NOT to Use This For

This command focuses on incomplete/missing/stub code. Use other commands for:

Security issues - Use /code:security for vulnerabilities, XSS, injection, auth bypass
Logic bugs - Use /code:logic for race conditions, type bugs, edge cases, control flow
Performance - Use /code:performance for slow queries, memory leaks, bundle size
Test gaps - Use /code:test for missing tests, weak assertions, test patterns
Compliance - Use /code:legal for GDPR, licensing, regulatory requirements

Ultradeep Mode

For high-stakes audits, use DEPTH=ultradeep to spawn each analyzer in a separate Claude Code tmux session:

# Show cost estimate before launching
/agileflow:code:completeness . DEPTH=ultradeep --dry-run
 
# Launch parallel sessions
/agileflow:code:completeness . DEPTH=ultradeep

The command will:

Show estimated cost and token usage
Confirm with you before spawning sessions
Monitor progress via sentinel files in docs/09-agents/ultradeep/
Collect all findings and run consensus
Generate final report

If tmux is unavailable, it automatically falls back to DEPTH=deep.

Command	Purpose
`/code:logic`	Find logic bugs and edge cases
`/code:security`	Detect security vulnerabilities
`/code:performance`	Find performance bottlenecks
`/code:test`	Analyze test coverage and quality
`/code:legal`	Check compliance and licensing
`/review`	AI-powered code review

/code:test /code:legal

On This Page

/code:completeness Quick Start Arguments How It Works Core Analyzers (Quick Mode - 5)Deep Mode Analyzers (2 Additional)Severity Scale Confidence Levels Examples Example Output Quick Depth Comparison What NOT to Use This For Ultradeep Mode Related Commands