Claude Code subagent imported from bigN13/ergoplanner_v1 (.claude/agents/README.md). Copyright stays with the author.

Claude Agents - Ergoplanner AI Suite

Overview

All agents are integrated with Task Master and enforce STRICT QUALITY GATES with zero tolerance for errors or warnings.

⚠️ CRITICAL: Quality Enforcement

NO agent can complete work if:

• ❌ Build errors exist
• ❌ Build warnings exist
• ❌ Tests are failing
• ❌ Code coverage < 80%
• ❌ Linting errors present

Agent Directory

Task Management Agents

Development Agents

Quality & Process Agents

Design Agents

Task Master File Locations

All agents reference these standard locations:

task_files:
  tasks: .taskmaster/tasks/tasks.json
  config: .taskmaster/config.json
  state: .taskmaster/state.json
  guide: .taskmaster/CLAUDE.md

project_files:
  standards: CLAUDE.md
  validation: scripts/validate-build.sh
  quality: scripts/run-quality-gates.sh
  pre_commit: scripts/pre-commit-check.sh

Quality Gate Scripts

All agents must use these validation scripts:

Primary Validation

# Before starting ANY work
./scripts/validate-build.sh

# After completing ANY task
./scripts/run-quality-gates.sh

# Before ANY commit
./scripts/pre-commit-check.sh

Continuous Validation

# Backend checks
dotnet build --no-incremental /warnaserror
dotnet test --collect:"XPlat Code Coverage"

# Frontend checks
npm run build
npm run lint -- --max-warnings 0
npm test -- --coverage --watchAll=false

# TypeScript
npx tsc --noEmit --strict

Agent Workflow Integration

1. Task Orchestration Flow

graph LR
    A[task-orchestrator] --> B[Analyze .taskmaster/tasks/tasks.json]
    B --> C[Run validate-build.sh]
    C --> D[Deploy task-executor agents]
    D --> E[Monitor quality gates]
    E --> F[Update task status]

2. Task Execution Flow

graph LR
    A[task-executor] --> B[Get task from tasks.json]
    B --> C[Implement with validation]
    C --> D[Run quality gates]
    D --> E[Update task status]

3. Task Verification Flow

graph LR
    A[task-checker] --> B[Review implementation]
    B --> C[Run all tests]
    C --> D[Verify coverage >= 80%]
    D --> E[Approve or reject]

Task Master Commands Used by Agents

# Task retrieval
task-master list                    # Get all tasks
task-master show <id>              # Get task details
task-master next                   # Get next priority task

# Status updates
task-master set-status --id=<id> --status=<status>
# Status values: pending, in-progress, review, blocked, done

# Progress tracking
task-master update-task --id=<id> --prompt="notes"
task-master update-subtask --id=<id> --prompt="progress"

# Validation
task-master validate-dependencies

Quality Enforcement Rules

Blocking Conditions

Agents MUST STOP work when:

• 🚨 Build error detected
• 🚨 Build warning present
• 🚨 Test failure
• 🚨 Coverage < 80%
• 🚨 Linting error
• 🚨 Security vulnerability

Task Completion Requirements

Agents can ONLY mark tasks complete when:

• ✅ Zero build errors
• ✅ Zero build warnings
• ✅ 100% tests passing
• ✅ >= 80% code coverage
• ✅ Zero linting issues
• ✅ Documentation updated
• ✅ Quality gates passed

Agent Communication Protocol

Task Assignment Format

task_assignment:
  task_id: [from tasks.json]
  objective: [clear goal]
  requirements:
    - quality_gates: ./scripts/validate-build.sh
    - coverage_minimum: 80%
    - zero_warnings: true
  validation:
    - pre_work: validate-build.sh
    - during_work: continuous
    - post_work: run-quality-gates.sh

Progress Reporting Format

progress_report:
  task_id: [id]
  status: [in-progress|blocked|review|done]
  quality_metrics:
    build: [pass|fail]
    tests: [pass|fail]
    coverage: [percentage]
    warnings: [count - must be 0]
  issues: [any blockers]
  next_steps: [planned actions]

Enhanced Agents

For agents with enhanced quality enforcement, use these versions:

• task-orchestrator-enhanced.md - Orchestrator with zero-tolerance
• task-executor-enhanced.md - Executor with strict validation

⛔ Forbidden Practices

Agents MUST NEVER:

• ❌ Skip validation to save time
• ❌ Commit with warnings
• ❌ Mark tasks done with failures
• ❌ Lower quality thresholds
• ❌ Use --no-verify flags
• ❌ Suppress or ignore errors
• ❌ Comment out failing tests

Success Metrics

Agent performance measured by:

1. Task completion rate with quality
2. Zero defect rate
3. Coverage maintenance (>= 80%)
4. Build success rate (must be 100%)
5. Time to quality gate passage

Quick Reference

Check Everything

./scripts/run-quality-gates.sh

Task Workflow

# Get task
task-master show <id>

# Start work (only if clean)
./scripts/validate-build.sh
task-master set-status --id=<id> --status=in-progress

# Complete (only if quality passes)
./scripts/run-quality-gates.sh
task-master set-status --id=<id> --status=done

Remember: Quality is NON-NEGOTIABLE across all agents!