tmux-team-creator
This skill should be used when users want to create a multi-agent tmux team for their project. It provides battle-tested templates for setting up autonomous AI agent teams that collaborate via tmux. Use this skill when users ask to create AI teams, set up multi-agent workflows, or build autonomous coding teams. The skill includes 4 sample team templates (2 software development + 2 research) that can be customized for any domain.
What this skill does
# Tmux Team Creator
## Overview
This skill enables creating powerful multi-agent AI teams that run autonomously in tmux sessions. The architecture is battle-tested on complex projects and can be customized for any domain.
**Key Insight**: Sample teams are TEMPLATES, not exact copies. When creating a team, customize roles, prompts, and workflows for the specific project.
---
## ⚠️ CRITICAL: Tmux Pane Detection Bug (EXTREMELY COMMON)
**THIS BUG WILL WASTE HOURS IF NOT PREVENTED!**
### The Bug
When agents need to determine which tmux pane they're running in, **NEVER use `tmux display-message -p '#{pane_index}'`** - this command returns the **ACTIVE/FOCUSED pane** (where the user's cursor is), NOT the pane where the agent is actually running!
### The Fix
**Always use the `$TMUX_PANE` environment variable:**
```bash
# WRONG - Returns active cursor pane, not your pane
tmux display-message -p '#{pane_index}'
# CORRECT - Returns YOUR actual pane
echo $TMUX_PANE
# Then look up this pane ID to get your role
tmux list-panes -a -F '#{pane_id} #{pane_index} #{@role_name}' | grep $TMUX_PANE
```
### Why This Matters
- In multi-agent teams, each pane has a specific role (PO, TL, DEV, etc.)
- Messages must route correctly based on pane roles
- If agents misidentify their pane, they send messages to wrong agents
- This causes hours of debugging "why is PO acting like DEV?"
### Where to Check
When creating a new team from templates, verify these files use `$TMUX_PANE`:
1. **Role prompt files** (`prompts/*_PROMPT.md`) - Should document correct pane detection
2. **Init commands** (`commands/init-role.md`) - Should use `$TMUX_PANE` for role detection
3. **Setup scripts** (`setup-team.sh`) - Should set `@role_name` on correct panes
4. **Hook scripts** (`hooks/*.sh`, `hooks/*.py`) - **CRITICAL:** Must use `$TMUX_PANE`
- **Bash hooks:** `ROLE=$(tmux show-options -t "$TMUX_PANE" -qv @role_name)`
- **Python hooks:** `tmux_pane = os.environ.get("TMUX_PANE")` then `["tmux", "show-options", "-pt", tmux_pane, "-qv", "@role_name"]`
- **WRONG:** `tmux show-options -pv @role_name` (queries cursor pane!)
- **FIX APPLIED:** `hooks/session_start_team_docs.py` (2026-01-02) - All new teams now use correct detection
### Prevention Checklist
When creating a new team, add this to ALL role prompts:
```markdown
## Tmux Pane Configuration & Role Detection
**CRITICAL: Correct Pane Detection**
**NEVER use `tmux display-message -p '#{pane_index}'`** - it returns the active/focused pane, not YOUR pane!
**Always use $TMUX_PANE environment variable:**
\`\`\`bash
# Find YOUR actual pane ID
echo "My pane: $TMUX_PANE"
# Look up your pane's role
tmux list-panes -a -F '#{pane_id} #{pane_index} #{@role_name}' | grep $TMUX_PANE
\`\`\`
```
**This bug has already been fixed in all sample team templates.** When creating new teams, copy the corrected patterns.
---
## FIRST: Ask User Which Team Template
**Before creating any team, ask the user which template they want to use.**
### Available Templates
#### Software Development Teams
| Template | Best For | Roles | Key Features |
|----------|----------|-------|--------------|
| **scrum-team** | Standard Scrum projects | PO, SM, TL, BE, FE, QA | Full Scrum framework, SM owns process improvement, Sprint-based |
| **scrum-minimal-team** | Small projects, MVPs | PO, SM, EX | Lightweight 3-person Scrum, EX combines TL+DEV+QA |
| **game-dev-team** | Game development projects | PM, GD, FE, BE, QA | Game-focused with design→implementation flow |
#### Research & Analysis Teams
| Template | Best For | Roles | Key Features |
|----------|----------|-------|--------------|
| **mckinsey-research-team** | Market research, competitive analysis | EM, RL, PR, SR, DA, QR | McKinsey 7-step methodology, MECE structuring, Pyramid Principle |
| **pg-insights-team** | Consumer insights, brand strategy | IM, MR, IA, SL, QR | P&G Three-Step Formula, human-centric, goosebumps test |
### Selection Logic
**If user specifies:**
- "Scrum team" / "standard Scrum" / "with Scrum Master" → Use `scrum-team`
- "minimal Scrum" / "small team" / "solo dev" / "MVP" / "lightweight" → Use `scrum-minimal-team`
- "game" / "game development" / "game project" → Use `game-dev-team`
- "market research" / "competitive analysis" / "McKinsey" / "research team" → Use `mckinsey-research-team`
- "consumer insights" / "P&G" / "brand strategy" / "emotional research" → Use `pg-insights-team`
**If user doesn't specify:**
Ask: "Which team template would you like to use?"
**Software Development:**
- **scrum-team** (Recommended for dev) - Full Scrum with PO, SM, TL, BE, FE, QA. SM owns process improvement.
- **game-dev-team** - Game development with PM, GD, FE, BE, QA. Design-first workflow.
**Research & Analysis:**
- **mckinsey-research-team** - McKinsey-style research with EM, RL, PR, SR, DA, QR. Hypothesis-driven, MECE structured.
- **pg-insights-team** - P&G-style consumer insights with IM, MR, IA, SL, QR. Human-centric, emotional + logical.
### Template Descriptions
#### 1. scrum-team (Recommended for most projects)
**Roles:** PO (Product Owner), SM (Scrum Master), TL (Tech Lead), BE (Backend), FE (Frontend), QA (Tester)
**Key Features:**
- Full Scrum framework adapted for AI teams
- SM owns process improvement with 4-checkpoint monitoring mechanism
- Sprint-based workflow with Planning, Review, Retrospective
- Black-box QA testing
- Prompt hygiene rules (add after 2-3 recurring issues, remove when learned)
**Best for:** Teams that want structured improvement, multiple sprints, quality focus
#### 2. scrum-minimal-team (Lightweight Scrum)
**Roles:** PO (Product Owner), SM (Scrum Master), EX (Executive = TL + DEV + QA)
**Key Features:**
- Minimal overhead for small projects
- EX combines architecture, development, and testing
- Same Sprint-based workflow as full Scrum
- SM still owns process improvement
- 3-pane tmux layout
**Best for:** Solo developers, MVPs, prototypes, personal projects wanting Scrum structure
#### 3. game-dev-team (Game Development)
**Roles:** DS (Game Designer), SM (Scrum Master), AR (Game Architect), DV (Game Developer), QA (Game QA)
**Key Features:**
- BMGD (BMAD Game Development) methodology + Scrum practices
- Design→Architecture→Implementation→Testing flow
- 60fps is non-negotiable - performance is a feature
- Playable increments every Sprint
- Design from player experience first
**Workflow:**
1. DS: Create Game Brief and GDD (mechanics, systems, content)
2. AR: Select engine, plan architecture, define performance budgets
3. DV: Implement Sprint stories with TDD
4. QA: Automated tests, playtests, performance profiling
**Best for:** Game development projects, interactive applications, real-time simulations
#### 4. mckinsey-research-team (McKinsey-style research)
**Roles:** EM (Engagement Manager), RL (Research Lead), PR (Primary Researcher), SR (Secondary Researcher), DA (Data Analyst), QR (Quality Reviewer)
**Key Features:**
- McKinsey 7-step problem-solving process
- MECE structuring (Mutually Exclusive, Collectively Exhaustive)
- Pyramid Principle for communication (lead with answer)
- Triangulation (multiple sources for key findings)
- EM owns process improvement with 4-checkpoint monitoring
**Workflow:**
1. Define Problem (EM ↔ Client)
2. Structure Problem (RL - MECE issue tree)
3. Prioritize Issues (EM + RL)
4. Plan Analysis (EM)
5. Conduct Analysis (PR, SR, DA in parallel)
6. Synthesize Findings (RL)
7. Communicate Recommendations (EM + RL → QR → Client)
**Best for:** Market research, competitive analysis, industry analysis, due diligence, strategy research
#### 5. pg-insights-team (P&G-style consumer insights)
**Roles:** IM (Insights Manager), MR (Moments Researcher), IA (Insight Analyst), SL (Strategy Lead), QR (Quality Reviewer)
**Key Features:**
- P&G Three-Step Insights Formula
- Human-centric research (everyday moments)
- Logic + Emotion connection
- Goosebumps test for insight validation
- IM owns process iRelated in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.