review-collector

Comprehensive validation of collector bot packages using 8 parallel specialized agents.

Usage

/review-collector [path]

Arguments:

• path (optional) - Path to collector package directory. Defaults to current directory.

Examples:

/review-collector package/amazon/aws/iam
/review-collector .

Description

This skill performs comprehensive validation of a collector bot package by spawning 8 specialized validation agents in PARALLEL. Each agent focuses on a specific aspect and reports violations with specific file:line references and fix suggestions.

Execution Instructions for Claude

When this skill is invoked:

Step 1: Determine Collector Path

Parse the args parameter. If empty, use current working directory.

Step 2: Spawn 8 Validation Agents in Parallel

Use the Task tool to spawn ALL 8 agents in a SINGLE message (parallel execution):

For each agent, read the corresponding instruction file from .claude/agents/ and use it as the agent prompt:

1. validate-config - Read .claude/agents/validate-config.md
2. validate-schema - Read .claude/agents/validate-schema.md
3. validate-mapping - Read .claude/agents/validate-mapping.md
4. validate-pagination - Read .claude/agents/validate-pagination.md
5. validate-groupid - Read .claude/agents/validate-groupid.md
6. validate-security - Read .claude/agents/validate-security.md
7. validate-performance - Read .claude/agents/validate-performance.md
8. validate-dependencies - Read .claude/agents/validate-dependencies.md

CRITICAL: Spawn all 8 in a single message with 8 Task tool calls for true parallel execution.

Pass to each agent:

[Agent Instructions from .md file]

---

TASK: Validate collector at: {absolutePath}

Follow the validation checklist above.

Return results in this EXACT JSON format:
{
  "category": "{AgentName}",
  "status": "PASS | FAIL | WARN",
  "issues": [
    {
      "file": "relative/path/from/collector/root",
      "line": number,
      "message": "Description",
      "severity": "error | warning | info",
      "suggestion": "How to fix"
    }
  ]
}

Wrap JSON in ```json code block.

Step 3: Wait for All Agents to Complete

Collect results from all 8 agents.

Step 4: Aggregate and Display Results

Parse JSON from each agent and display:

=================================================================================
🔍 Collector Review Results
=================================================================================

✅ Configuration: PASS
   📄 8 files checked, 0 issues

✅ Schema: PASS
   📄 3 classes validated, all implemented correctly

❌ Mapping: FAIL (3 issues)
   ❌ src/Mappers.ts:45 - Nested mapping detected (should be flat)
      💡 Use group IDs instead: groups: source.groupIds.sort()

   ❌ src/Mappers.ts:67 - Date field using DateTime format
      💡 Use Object.assign workaround - see ADVANCED_MAPPING_GUIDE.md

   ⚠️  src/Mappers.ts:89 - Enum case mismatch
      💡 Change 'ACTIVE' to 'Active'

✅ Pagination: PASS
   📄 All pages processed, pageSize optimized

⚠️  GroupId: WARN (1 issue)
   ⚠️  src/CollectorImpl.ts:124 - GroupId pattern unclear
       💡 Consider: ${accountId}-${region} for regional resources

✅ Security: PASS
   📄 No vulnerabilities detected

✅ Performance: PASS
   📄 Concurrency within limits, memory managed well

❌ Dependencies: FAIL (1 issue)
   ❌ package.json:38 - Product package in dependencies
      💡 Remove @auditlogic/product-amazon-aws-iam (transitive dependency)

=================================================================================
📊 Overall Results
=================================================================================
   ✅ Pass: 6
   ❌ Fail: 2
   ⚠️  Warn: 1
   📈 Compliance: 75%

⚠️  Action required: Fix 2 critical issue(s) before commit.
=================================================================================

Step 5: Provide Summary

• If all PASS → "🎉 All validations passed! Collector is ready for commit."
• If any FAIL → "⚠️ Action required: Fix N critical issues."
• If only WARN → "✓ Ready with warnings: Review N warnings."

Compliance score: (pass_count / 8) × 100

Notes

• All agents run in parallel for speed
• Each agent is independent
• All agent instructions are in .claude/agents/*.md
• Results are structured JSON for consistent parsing
• File paths are relative to collector root
• Suggestions are actionable and reference documentation