ACE Status

Display comprehensive statistics about your ACE playbook.

Call ace-cli to get current statistics:

#!/usr/bin/env bash
set -euo pipefail

# CLI detection: ace-cli (preferred) or ce-ace (fallback)
if ! command -v ace-cli >/dev/null 2>&1; then
  echo "❌ ace-cli not found - Install: npm install -g @ace-sdk/cli"
  exit 1
fi

# Check for jq (required for JSON formatting)
if ! command -v jq >/dev/null 2>&1; then
  echo "❌ jq not found - Install: brew install jq (macOS) or apt-get install jq (Linux)"
  exit 1
fi

# Version check: Query npm registry directly (Issue #8 - CLI version check broken)
INSTALLED_VERSION=$(ace-cli --version 2>/dev/null | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1 || echo "unknown")
LATEST_VERSION=$(curl -s "https://registry.npmjs.org/@ace-sdk/cli/latest" 2>/dev/null | jq -r '.version // "unknown"' || echo "unknown")

if [ "$INSTALLED_VERSION" != "unknown" ] && [ "$LATEST_VERSION" != "unknown" ]; then
  if [ "$INSTALLED_VERSION" != "$LATEST_VERSION" ]; then
    echo "💡 Update available: ace-cli v$INSTALLED_VERSION → v$LATEST_VERSION"
    echo "   Run: npm install -g @ace-sdk/cli@latest"
    echo ""
  fi
fi

# Optional: Export environment variables if available from .claude/settings.json
# ace-cli will use these if provided, otherwise fallback to global config
export ACE_ORG_ID="${ACE_ORG_ID:-}"
export ACE_PROJECT_ID="${ACE_PROJECT_ID:-}"

# Run ace-cli status and capture output
# Filter out CLI update notifications (💡 lines) that break JSON parsing
RAW_OUTPUT=$(ace-cli status --json 2>&1)
EXIT_CODE=$?
STATUS_OUTPUT=$(echo "$RAW_OUTPUT" | grep -v '^💡' | grep -v '^$')

# Check if command succeeded
if [ $EXIT_CODE -ne 0 ]; then
  echo "❌ Failed to get ACE status"
  echo ""
  echo "Error details:"
  echo "$STATUS_OUTPUT"
  echo ""
  echo "Common fixes:"
  echo "  1. Run: /ace-configure to setup configuration"
  echo "  2. Verify global config exists: cat ~/.config/ace/config.json"
  echo "  3. Check API token is valid at: https://ace.code-engine.app/settings"
  exit 1
fi

# Verify we got valid JSON
if ! echo "$STATUS_OUTPUT" | jq empty 2>/dev/null; then
  echo "❌ Invalid response from ace-cli (not valid JSON)"
  echo ""
  echo "Response:"
  echo "$STATUS_OUTPUT"
  exit 1
fi

# Format output for readability
echo "$STATUS_OUTPUT" | jq -r '
  "📊 ACE Playbook Status",
  "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━",
  "Organization: \(.org_id // "Not configured")",
  "Project: \(.project_id // "Not configured")",
  "",
  "📚 Total Patterns: \(.total_bullets // 0)",
  "",
  "By Section:",
  "  • Strategies & Rules: \(.by_section.strategies_and_hard_rules // 0)",
  "  • Code Snippets: \(.by_section.useful_code_snippets // 0)",
  "  • Troubleshooting: \(.by_section.troubleshooting_and_pitfalls // 0)",
  "  • APIs to Use: \(.by_section.apis_to_use // 0)",
  "",
  "📈 Average Confidence: \((.avg_confidence // 0) * 100 | floor)%"
'

What You'll See

Playbook Summary:

Total bullets across all sections
Bullets by section (strategies, snippets, troubleshooting, APIs)
Average confidence score

Top Helpful Bullets:

5 most helpful bullets (highest ✅ counts)
Shows which patterns are most valuable

Top Harmful Bullets:

5 most harmful bullets (highest ❌ counts)
Shows which patterns are misleading

Example Output

{
  "total_bullets": 42,
  "by_section": {
    "strategies_and_hard_rules": 10,
    "useful_code_snippets": 15,
    "troubleshooting_and_pitfalls": 12,
    "apis_to_use": 5
  },
  "avg_confidence": 0.78,
  "top_helpful": [
    {
      "id": "ctx-1737387600-a1b2c",
      "section": "strategies_and_hard_rules",
      "content": "Always verify npm package names...",
      "helpful": 12,
      "harmful": 0,
      "confidence": 1.0
    },
    ...
  ],
  "top_harmful": [...]
}

How to Interpret

High helpful count → Pattern proven useful across multiple tasks High harmful count → Pattern caused errors, needs revision High confidence → helpful/(helpful+harmful) ratio is strong Low confidence → Pattern unreliable, may be pruned at 0.30 threshold

Storage

ACE v3.0 stores playbooks in:

Remote: ACE Storage Server (FastAPI + ChromaDB)
Project-specific: Isolated by project ID
Multi-tenant: Your org's patterns are private

Next Steps

After checking status:

If playbook is empty (0 bullets):
- Run /ace:bootstrap to populate initial patterns from git/docs/code
- Or start coding - ACE will learn automatically!
If not configured yet:
- Run /ace:configure to set up ACE server connection
If ACE not initialized in project:
- Run /ace:claude-init to add ACE instructions to CLAUDE.md

/ace-status