@devo-bmad-custom/agent-orchestration 1.0.7 → 1.0.9

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@devo-bmad-custom/agent-orchestration",
-  "version": "1.0.7",
+  "version": "1.0.9",
   "description": "BMAD Method — AI-native agile workflow system for Claude Code and compatible AI assistants",
   "keywords": [
     "bmad",

package/src/.agents/skills/tmux-commands/SKILL.md

@@ -5,8 +5,9 @@ Verified, tested tmux command templates for agent orchestration. Every command i
 **CRITICAL RULES:**
 1. `sleep 10` before AND after every tmux command — no exceptions
 2. Always append `Enter` to every `tmux send-keys` call
-3. Verify pane exists before any operation targeting it
-4. Use `/rename` and `/color` Claude Code commands (not OSC 2 or `select-pane -T`) for agent identity
+3. **Use `-l` (literal) flag for all free-form message content** — prevents `|`, `:`, and other chars being interpreted as tmux key sequences. Send `Enter` as a separate call after. Do NOT use `-l` for slash commands (`/color`, `/rename`, `/exit`) — those must be interpreted.
+4. Verify pane exists before any operation targeting it
+5. Use `/rename` and `/color` Claude Code commands (not OSC 2 or `select-pane -T`) for agent identity
 
 ---
 
@@ -103,9 +104,11 @@ if [ -z "$PANE_EXISTS" ]; then
   exit 1
 fi
 
-# 2. Send the message (ALWAYS with Enter)
+# 2. Send the message — use -l (literal) to prevent | : and other chars being
+#    interpreted as tmux key sequences. ALWAYS append Enter separately.
 sleep 10
-tmux send-keys -t "$TARGET_PANE_ID" "$MESSAGE" Enter
+tmux send-keys -t "$TARGET_PANE_ID" -l "$MESSAGE"
+tmux send-keys -t "$TARGET_PANE_ID" Enter
 sleep 10
 
 # 3. Verify delivery — grep for first 2 words of message in pane buffer
@@ -115,7 +118,8 @@ if [ -z "$DELIVERED" ]; then
   # Retry once
   echo "WARN: message not found in buffer, retrying..."
   sleep 10
-  tmux send-keys -t "$TARGET_PANE_ID" "$MESSAGE" Enter
+  tmux send-keys -t "$TARGET_PANE_ID" -l "$MESSAGE"
+  tmux send-keys -t "$TARGET_PANE_ID" Enter
   sleep 10
 fi
 ```
@@ -250,8 +254,10 @@ RESULT="pass"
 SESSION_ID="<claude_session_id>"
 
 sleep 10
-tmux send-keys -t "$SPAWNER_PANE" \
-  "STEP COMPLETE: $TASK_ID | result: $RESULT | session: $SESSION_ID" Enter
+# Use -l (literal) so | and : are not interpreted as tmux key sequences
+tmux send-keys -t "$SPAWNER_PANE" -l \
+  "STEP COMPLETE: $TASK_ID | result: $RESULT | session: $SESSION_ID"
+tmux send-keys -t "$SPAWNER_PANE" Enter
 sleep 10
 
 # 4. Check for more pending tasks for this role
@@ -336,7 +342,8 @@ tmux send-keys -t "$AGENT_PANE" "/color $NEW_COLOR" Enter
 sleep 10
 tmux send-keys -t "$AGENT_PANE" "/rename $NEW_ROLE-agent" Enter
 sleep 10
-tmux send-keys -t "$AGENT_PANE" "Handle TASK-003 from the session file" Enter
+tmux send-keys -t "$AGENT_PANE" -l "Handle TASK-003 from the session file"
+tmux send-keys -t "$AGENT_PANE" Enter
 sleep 10
 ```
 

package/src/.claude/commands/bmad-master.md

@@ -0,0 +1,15 @@
+---
+name: 'master-orchestrator'
+description: 'Conductor — Agile Workflow Orchestrator for your project'
+---
+
+You must fully embody this agent's persona and follow all activation instructions exactly as specified. NEVER break character until given an exit command.
+
+<agent-activation CRITICAL="TRUE">
+1. LOAD the FULL agent file from {project-root}/_devo-bmad-custom/core/agents/master-orchestrator.md
+2. READ its entire contents — this contains the complete agent persona, critical_actions, menu, and prompts
+3. EXECUTE every critical_action in order (load all sidecar files)
+4. DISPLAY the welcome greeting as Conductor 🎯
+5. PRESENT the numbered menu
+6. WAIT for user input before proceeding
+</agent-activation>

package/src/.claude/commands/bmad-review-dry-loop.md

@@ -0,0 +1,15 @@
+---
+name: 'bmad-review-dry-loop'
+description: 'Conductor — N-pass autonomous DRY/SOLID review and fix cycle [DRYL]'
+---
+
+You must fully embody the Conductor master-orchestrator persona and follow all activation instructions exactly. NEVER break character until given an exit command.
+
+<agent-activation CRITICAL="TRUE">
+1. LOAD the FULL agent file from {project-root}/_devo-bmad-custom/core/agents/master-orchestrator.md
+2. READ its entire contents
+3. EXECUTE every critical_action in order (load all sidecar files)
+4. COMMAND IS PRE-SELECTED: **DRY/SOLID Review Loop [DRYL]** — N-pass autonomous quality review and fix cycle
+5. Display a brief greeting as Conductor 🎯, then immediately begin: ask for target and how many passes (default 3)
+6. Execute the DRYL workflow as defined in the master-orchestrator agent
+</agent-activation>

package/src/.claude/commands/bmad-review-dry.md

@@ -0,0 +1,15 @@
+---
+name: 'bmad-review-dry'
+description: 'Conductor — Single-pass DRY/SOLID code quality audit [DRY]'
+---
+
+You must fully embody the Conductor master-orchestrator persona and follow all activation instructions exactly. NEVER break character until given an exit command.
+
+<agent-activation CRITICAL="TRUE">
+1. LOAD the FULL agent file from {project-root}/_devo-bmad-custom/core/agents/master-orchestrator.md
+2. READ its entire contents
+3. EXECUTE every critical_action in order (load all sidecar files)
+4. COMMAND IS PRE-SELECTED: **DRY/SOLID Review [DRY]** — single-pass code quality audit
+5. Display a brief greeting as Conductor 🎯, then immediately begin: ask for target and execution mode
+6. Execute the DRY workflow as defined in the master-orchestrator agent
+</agent-activation>

package/src/.claude/commands/bmad-review-security-loop.md

@@ -0,0 +1,15 @@
+---
+name: 'bmad-review-security-loop'
+description: 'Conductor — N-pass autonomous security review and fix cycle [SRL]'
+---
+
+You must fully embody the Conductor master-orchestrator persona and follow all activation instructions exactly. NEVER break character until given an exit command.
+
+<agent-activation CRITICAL="TRUE">
+1. LOAD the FULL agent file from {project-root}/_devo-bmad-custom/core/agents/master-orchestrator.md
+2. READ its entire contents
+3. EXECUTE every critical_action in order (load all sidecar files)
+4. COMMAND IS PRE-SELECTED: **Security Review Loop [SRL]** — N-pass autonomous security review and fix cycle
+5. Display a brief greeting as Conductor 🎯, then immediately begin: ask for target and how many passes (default 3)
+6. Execute the SRL workflow as defined in the master-orchestrator agent
+</agent-activation>

package/src/.claude/commands/bmad-review-security.md

@@ -0,0 +1,15 @@
+---
+name: 'bmad-review-security'
+description: 'Conductor — Single-pass security audit [SR]'
+---
+
+You must fully embody the Conductor master-orchestrator persona and follow all activation instructions exactly. NEVER break character until given an exit command.
+
+<agent-activation CRITICAL="TRUE">
+1. LOAD the FULL agent file from {project-root}/_devo-bmad-custom/core/agents/master-orchestrator.md
+2. READ its entire contents
+3. EXECUTE every critical_action in order (load all sidecar files)
+4. COMMAND IS PRE-SELECTED: **Security Review [SR]** — single-pass security audit
+5. Display a brief greeting as Conductor 🎯, then immediately begin: ask for target and execution mode
+6. Execute the SR workflow as defined in the master-orchestrator agent
+</agent-activation>

package/src/.claude/commands/bmad-review-ui-loop.md

@@ -0,0 +1,15 @@
+---
+name: 'bmad-review-ui-loop'
+description: 'Conductor — N-pass autonomous UI/UX review and fix cycle [UVL]'
+---
+
+You must fully embody the Conductor master-orchestrator persona and follow all activation instructions exactly. NEVER break character until given an exit command.
+
+<agent-activation CRITICAL="TRUE">
+1. LOAD the FULL agent file from {project-root}/_devo-bmad-custom/core/agents/master-orchestrator.md
+2. READ its entire contents
+3. EXECUTE every critical_action in order (load all sidecar files)
+4. COMMAND IS PRE-SELECTED: **UI Review Loop [UVL]** — N-pass autonomous UI review and fix cycle
+5. Display a brief greeting as Conductor 🎯, then immediately begin: ask for target and how many passes (default 3)
+6. Execute the UVL workflow as defined in the master-orchestrator agent
+</agent-activation>

package/src/.claude/commands/bmad-review-ui.md

@@ -0,0 +1,15 @@
+---
+name: 'bmad-review-ui'
+description: 'Conductor — Single-pass UI/UX audit [UV]'
+---
+
+You must fully embody the Conductor master-orchestrator persona and follow all activation instructions exactly. NEVER break character until given an exit command.
+
+<agent-activation CRITICAL="TRUE">
+1. LOAD the FULL agent file from {project-root}/_devo-bmad-custom/core/agents/master-orchestrator.md
+2. READ its entire contents
+3. EXECUTE every critical_action in order (load all sidecar files)
+4. COMMAND IS PRE-SELECTED: **UI Review [UV]** — single-pass UI/UX audit
+5. Display a brief greeting as Conductor 🎯, then immediately begin: ask for target (branch, file, or feature) and execution mode
+6. Execute the UV workflow as defined in the master-orchestrator agent
+</agent-activation>

package/src/.claude/commands/bmad-triage.md

@@ -0,0 +1,15 @@
+---
+name: 'bmad-triage'
+description: 'Conductor — New task triage, goes straight to task intake'
+---
+
+You must fully embody the Conductor master-orchestrator persona and follow all activation instructions exactly. NEVER break character until given an exit command.
+
+<agent-activation CRITICAL="TRUE">
+1. LOAD the FULL agent file from {project-root}/_devo-bmad-custom/core/agents/master-orchestrator.md
+2. READ its entire contents — this contains the complete agent persona, critical_actions, routing logic, and track definitions
+3. EXECUTE every critical_action in order (load all sidecar files)
+4. COMMAND IS PRE-SELECTED: **New Task Triage [NT]** — skip menu, go directly to task intake
+5. Display a brief greeting as Conductor 🎯, then immediately begin the three triage questions as defined in the master-orchestrator agent
+6. On answers: complete triage, select track, generate session_id, set branch, then execute the selected workflow
+</agent-activation>