Initial commit

commands/monitor/recommend.md

@@ -0,0 +1,36 @@
+---
+name: monitor:recommend
+description: Get smart workflow and optimization recommendations based on learned patterns
+---
+
+EXECUTE THESE BASH COMMANDS DIRECTLY (no agents, no skills):
+
+First, find the plugin installation path:
+```bash
+PLUGIN_PATH=$(find ~/.claude -name "exec_plugin_script.py" 2>/dev/null | head -1 | sed 's|/lib/exec_plugin_script.py||')
+echo "Plugin found at: $PLUGIN_PATH"
+```
+
+Step 1 - Generate recommendations using project patterns:
+```bash
+python3 "$PLUGIN_PATH/lib/exec_plugin_script.py" recommendation_engine.py --dir ./.claude-patterns --task "your task description here"
+```
+
+Step 2 - For general recommendations (no specific task):
+```bash
+python3 "$PLUGIN_PATH/lib/exec_plugin_script.py" recommendation_engine.py --dir ./.claude-patterns
+```
+
+Step 3 - For JSON output (programmatic use):
+```bash
+python3 "$PLUGIN_PATH/lib/exec_plugin_script.py" recommendation_engine.py --dir ./.claude-patterns --task "your task" --format json
+```
+
+Report will show:
+- Recommended approach with confidence level
+- Quality predictions and time estimates
+- Skill suggestions with success rates
+- Risk assessment with mitigations
+- Alternative approaches with trade-offs
+
+The recommendations will use patterns from your current project directory at ./.claude-patterns/