Step 1: Discover Target Logs

• Invoke log-lister skill with filters
• Get list of log paths to validate
• Report total count: "Found N logs to validate"

Step 2: Initialize Worker Pool

Execute scripts/init-worker-pool.sh:

• Create N worker processes (default: 5)
• Set up job queue
• Initialize progress tracking

Step 3: Queue Validation Jobs

For each log path:

• Add validation job to queue
• Job: invoke log-validator with log_path

Step 4: Execute in Parallel

Execute scripts/parallel-execute.sh:

• Workers consume jobs from queue
• Each worker validates one log at a time
• Results collected in shared storage (flock for concurrency)
• Progress reported: "Validated 25/100 logs..."

Step 5: Aggregate Results

Execute scripts/aggregate-results.sh:

• Collect all validation results
• Count: passed, failed (by severity), warnings
• Group errors by type
• Identify most common issues

Step 6: Return Summary

{
  "operation": "batch-validate",
  "status": "completed",
  "total_logs": 100,
  "results": {
    "passed": 85,
    "failed": 10,
    "warnings": 5
  },
  "failures": [
    {
      "log_path": ".fractary/logs/test/test-042.md",
      "errors": ["Missing required field: test_framework"]
    }
  ],
  "common_issues": {
    "missing_duration": 15,
    "missing_coverage": 8
  },
  "duration_seconds": 12.5,
  "parallel_workers": 5
}

Batch Operation: batch-archive

Step 1: Discover Archival Candidates

• Invoke log-lister skill
• For each log, check retention policy:
  • Load types/{log_type}/retention-config.json
  • Calculate retention expiry date
  • Mark logs past retention period

Step 2: Validate Before Archive (unless skipped)

• Optionally validate all candidates
• Skip logs with critical errors

Step 3: Update Status in Parallel

For each archival candidate:

• Update frontmatter status to "archived"
• Add archive_date field
• Preserve in current location (for now)

Step 4: Return Archive Report

{
  "operation": "batch-archive",
  "total_candidates": 45,
  "archived": 42,
  "skipped": 3,
  "skipped_reasons": {
    "validation_failed": 2,
    "retention_not_expired": 1
  },
  "by_type": {
    "session": 15,
    "build": 18,
    "test": 9
  }
}

Batch Operation: batch-reclassify

Step 1: Find _untyped Logs

• List all logs with log_type="_untyped"
• Report count

Step 2: Classify in Parallel

For each _untyped log:

• Invoke log-classifier skill
• Get recommended type and confidence
• If confidence >= threshold, mark for reclassification

Step 3: Preview Reclassifications

Show user:

Reclassification Preview:
  - log-001.md: _untyped → test (95% confident)
  - log-002.md: _untyped → build (87% confident)
  - log-003.md: _untyped → operational (72% confident)
  - log-004.md: _untyped → [uncertain, keeping as _untyped] (45%)

Step 4: Apply (if auto_apply or confirmed)

For each approved reclassification:

• Update frontmatter log_type
• Move to correct type directory
• Revalidate against new schema

Step 5: Return Summary

{
  "operation": "batch-reclassify",
  "total_untyped": 50,
  "reclassified": 35,
  "uncertain": 10,
  "failed": 5,
  "reclassifications": {
    "test": 15,
    "build": 10,
    "session": 5,
    "operational": 5
  }
}

Batch Operation: batch-cleanup

Step 1: Find Expired Logs (with buffer)

• List archived logs
• Check retention expiry + buffer period
• Identify deletion candidates

Step 2: Safety Check

• REQUIRE dry_run=false for actual deletion
• REQUIRE explicit confirmation for production/critical types
• NEVER delete audit logs (per retention policy)

Step 3: Delete (if authorized)

• Remove log files
• Update archive index
• Log deletions to audit trail

Step 4: Return Cleanup Report

{
  "operation": "batch-cleanup",
  "dry_run": false,
  "total_candidates": 120,
  "deleted": 115,
  "protected": 5,
  "protected_reasons": {
    "audit_never_delete": 3,
    "production_deployment": 2
  },
  "space_freed_mb": 45
}