Decision Guide: When to Use What
A comprehensive guide to help you choose the right tools, commands, agents, and approaches in Claude Copilot.
Note: This is the complete reference. For a quick overview integrated into Claude Code's instructions, see CLAUDE.md.
Feature Selection
Feature Comparison
| Feature |
Invocation |
Persistence |
Best For |
When NOT to Use |
| Memory |
Auto |
Cross-session |
Context preservation, decisions, lessons |
Short-term notes, temporary data |
| Agents |
Protocol |
Session |
Expert tasks, complex work |
Simple commands, quick tasks |
| Skills |
Auto |
On-demand |
Reusable patterns, workflows |
One-off solutions |
| Commands |
Manual |
Session |
Quick shortcuts, workflows |
Complex multi-step processes |
| Extensions |
Auto |
Permanent |
Team standards, custom methodologies |
Personal preferences |
Memory vs Skills vs Extensions
| Question |
Use Memory |
Use Skills |
Use Extensions |
| Does it change per project? |
✓ |
|
|
| Is it a one-time decision? |
✓ |
|
|
| Is it a reusable pattern? |
|
✓ |
|
| Does the whole team need it? |
|
|
✓ |
| Is it company-specific? |
|
|
✓ |
| Does it override base behavior? |
|
|
✓ |
| Is it a lesson learned? |
✓ |
|
|
| Is it a workflow/automation? |
|
✓ |
|
Command Selection
Command Decision Matrix
| I want to... |
Command |
When |
Where to Run |
| Set up Claude Copilot first time |
/setup |
Once per machine |
~/.claude/copilot |
| Add Copilot to new project |
/setup-project |
Once per project |
Project root |
| Update project files |
/update-project |
After framework updates |
Project root |
| Update framework itself |
/update-copilot |
When new version available |
Any directory |
| Create team knowledge |
/knowledge-copilot |
Once per team/company |
Any directory |
| Start fresh work |
/protocol [task] |
Each work session |
Project root |
| Resume previous work |
/continue [stream] |
When returning to work |
Project root |
| Run parallel work streams |
/orchestrate start |
Set up worktrees, launch via Task tool |
Project root |
| Monitor orchestration |
/orchestrate status |
During parallel execution |
Project root |
| Merge completed streams |
/orchestrate merge |
After streams complete |
Project root |
| Verify MCP servers |
/mcp |
After setup, troubleshooting |
Project root |
Command Arguments:
/protocol <task> - Auto-detect task type and route to agent (e.g., /protocol fix the login bug)
/continue <stream> - Resume specific parallel stream (e.g., /continue Stream-B)
Setup Command Flowchart
Are you using Claude Copilot for the first time?
├─ YES → Run /setup in ~/.claude/copilot
│ ↓
│ Is this a new project?
│ ├─ YES → Run /setup-project in project
│ └─ NO → Run /update-project in existing project
│
└─ NO → Did you update Claude Copilot?
├─ YES → Run /update-project in all projects
└─ NO → Just use /protocol or /continue to work
Agent Selection
Agent Routing Matrix
| Task Type |
Primary Agent |
Secondary Agent(s) |
Why This Flow |
| Bug Fix |
qa |
→ me |
QA reproduces, Engineer fixes |
| New Feature |
sd |
→ uxd → uid |
Service → UX → Implementation |
| API Design |
ta |
→ me → doc |
Architecture → Code → Docs |
| Security Review |
sec |
|
Security expertise |
| Performance Issue |
ta |
→ me |
Design analysis → Implementation |
| UI Component |
uids |
→ uid |
Visual design → Code |
| Documentation |
doc |
|
Technical writing |
| Deployment |
do |
|
DevOps expertise |
| Architecture Decision |
ta |
|
System design |
| User Research |
sd |
|
Experience strategy |
| Copy/Messaging |
cw |
|
Content writing |
Scenario-Based Agent Selection
| Scenario |
Start With |
Reasoning |
| "Users can't login" |
/protocol (DEFECT) → @agent-qa |
Needs reproduction and diagnosis |
| "Add dark mode" |
/protocol (EXPERIENCE) → @agent-sd |
Experience change requires journey analysis |
| "Optimize database queries" |
/protocol (FEATURE) → @agent-ta |
Architecture-level optimization |
| "Deploy to production" |
/protocol (DEVOPS) → @agent-do |
Infrastructure task |
| "Security audit" |
/protocol (SECURITY) → @agent-sec |
Security expertise required |
| "Write API docs" |
/protocol (DOCUMENTATION) → @agent-doc |
Documentation specialist |
| "Refactor auth module" |
/protocol (ARCHITECTURE) → @agent-ta |
Design decision needed |
| "Fix button alignment" |
/protocol (DEFECT) → @agent-uid |
UI implementation fix |
Extension vs Override vs Skills
Extension Type Decision Tree
Do you need to customize agent behavior?
├─ NO → Use base agents as-is
│
└─ YES → Do you want to replace the entire agent?
├─ YES → Use OVERRIDE (.override.md)
│ Example: Completely custom methodology
│
└─ NO → Do you want to add/enhance sections?
├─ YES → Use EXTENSION (.extension.md)
│ Example: Add checklists, templates
│
└─ NO → Do you just need to inject skills?
└─ Use SKILLS (.skills.json)
Example: Company-specific tools
Extension Type Comparison
| Goal |
Extension Type |
File Pattern |
Scope |
Difficulty |
| Add company checklist |
extension |
agent.extension.md |
Section merge |
Easy |
| Replace methodology |
override |
agent.override.md |
Full replacement |
Hard |
| Inject team tools |
skills |
agent.skills.json |
Skill list only |
Medium |
| Enhance templates |
extension |
agent.extension.md |
Section merge |
Easy |
| Custom process flow |
override |
agent.override.md |
Full replacement |
Hard |
| Add domain knowledge |
skills |
agent.skills.json |
Skill list only |
Medium |
Work Session Decisions
Starting Work Decision Matrix
| Situation |
Use |
Why |
| Brand new task |
/protocol or /protocol <task> |
Classify and route to expert |
| Continuing yesterday's work |
/continue |
Load context from memory |
| Resume specific parallel stream |
/continue <stream> |
Jump directly to stream work |
| Quick question |
Just ask |
No need for protocol |
| Exploring ideas |
Just ask |
Protocol is for execution |
| Complex multi-step task |
/protocol |
Agent expertise needed |
| Resume after interruption |
/continue |
Context restoration |
Protocol vs Direct Conversation
| Use /protocol When... |
Use Direct Conversation When... |
| Building a feature |
Asking a question |
| Fixing a bug |
Exploring ideas |
| Making architecture changes |
Getting quick help |
| Need expert guidance |
Simple task |
| Multi-step process |
Single action |
| Want structured approach |
Want flexibility |
Knowledge Repository Decisions
When to Create Knowledge Repository
| Indicator |
Action |
| Team > 1 person |
Create knowledge repo |
| Company has style guide |
Document in knowledge repo |
| Custom methodologies |
Create agent extensions |
| Repeated explanations |
Document once in knowledge |
| Onboarding takes days |
Create knowledge repo |
| Inconsistent outputs |
Define standards in knowledge |
Knowledge vs Skills vs Memory
| Type of Information |
Store In |
Why |
| Company values |
Knowledge |
Permanent, team-wide |
| Voice/tone guidelines |
Knowledge |
Permanent, team-wide |
| Design system |
Knowledge |
Permanent, team-wide |
| Reusable code patterns |
Skills |
On-demand, reusable |
| Project decisions |
Memory |
Project-specific, temporal |
| Lessons learned |
Memory |
Project-specific, evolving |
| API standards |
Knowledge |
Team standard |
| Deployment process |
Skills |
Reusable workflow |
| Why we chose X |
Memory |
Project context |
Stream Management Decisions
When to Use Streams
| Scenario |
Use Streams? |
Why |
| Single developer, sequential work |
No |
Standard task flow is simpler |
| Multiple parallel work areas |
Yes |
Prevent file conflicts |
| Large feature with independent components |
Yes |
Parallel development |
| Bug fix while feature in progress |
Yes |
Isolate changes |
| Foundation → parallel → integration workflow |
Yes |
Enforce dependencies |
Stream Pattern
Foundation Phase:
- Stream-A: Core infrastructure that parallel work depends on
- Example: Schema changes, base types, shared utilities
Parallel Phase:
- Stream-B, Stream-C, Stream-D: Independent work streams
- Each touches different files
- Can run simultaneously in separate Claude Code sessions
- Example: Stream-B (API endpoints), Stream-C (UI components), Stream-D (tests)
Integration Phase:
- Stream-Z: Combines parallel streams
- Example: Documentation, final validation, release prep
Stream Tools Usage
| Command |
When to Use |
Example |
tc stream list --json |
View all streams in initiative |
See progress across parallel work |
tc stream get <id> --json |
Get detailed stream info |
Check Stream-B status before resuming |
git diff |
Before creating tasks |
Ensure no file conflicts with other streams |
Stream Metadata in Tasks
When creating tasks with streams, include:
metadata: {
streamId: "Stream-B", // Auto-generated ID
streamName: "command-updates", // Human-readable name
streamPhase: "parallel", // foundation | parallel | integration
files: [ // Files this task will modify
"path/to/file1.ts",
"path/to/file2.md"
],
streamDependencies: ["Stream-A"] // Must complete before this stream
}
Troubleshooting Decisions
When Something Goes Wrong
| Problem |
Check First |
Then Try |
Last Resort |
| Command not found |
Machine setup complete? |
/setup in ~/.claude/copilot |
Reinstall |
| MCP not connecting |
.mcp.json paths absolute? |
Rebuild servers |
Check Node version |
| Agent not routing |
Is task description clear? |
Rephrase request |
Use agent directly |
| Memory not persisting |
WORKSPACE_ID set? |
Check .mcp.json |
Rebuild memory server |
| Knowledge not found |
Symlink exists? |
/knowledge-copilot |
Manual link |
| Skills not loading |
Skills server running? |
/mcp to verify |
Check logs |
Quick Reference Tables
Installation Decision Matrix
| You Are... |
Steps Required |
Commands to Run |
| New solo user |
Clone → Setup → Project |
/setup → /setup-project |
| New team member |
Clone → Setup → Link Knowledge → Project |
/setup → /knowledge-copilot → /setup-project |
| Existing user, new project |
Project setup only |
/setup-project |
| Updating framework |
Update → Rebuild → Sync projects |
/update-copilot → /update-project |
Daily Workflow Matrix
| Beginning of Day |
During Day |
End of Day |
/continue to resume |
Work naturally with agents |
initiative_update to save progress |
Or /protocol for new task |
Route complex work to specialists |
Document decisions in memory |
Check /mcp if needed |
Use skills as needed |
Note lessons learned |
Best Practices Summary
Do This
| Context |
Best Practice |
| Starting work |
Use /continue to load context |
| Complex tasks |
Use /protocol to engage experts |
| Team standards |
Create knowledge repository |
| Reusable patterns |
Save as skills |
| Project decisions |
Store in memory with context |
| End of session |
Update initiative with progress |
Don't Do This
| Anti-Pattern |
Instead Do |
| Hardcode paths in extensions |
Use relative paths |
| Store secrets in knowledge |
Use environment variables |
| Override agents unnecessarily |
Use extensions first |
| Skip memory updates |
Update at end of session |
| Create skills for one-offs |
Just solve the problem |
| Ignore agent routing |
Trust the protocol |
Decision Flowchart: Complete Setup
START: Do you have Claude Copilot installed?
│
├─ NO → Clone to ~/.claude/copilot
│ ↓
│ Run /setup in ~/.claude/copilot
│ ↓
│ Are you part of a team?
│ ├─ YES → Clone team knowledge repo
│ │ Run /knowledge-copilot to link
│ │ ↓
│ └─ NO → Do you want to create team knowledge?
│ ├─ YES → Run /knowledge-copilot
│ └─ NO → Skip knowledge setup
│ ↓
│ Run /setup-project in your project
│ ↓
│ Done! Use /protocol to start working
│
└─ YES → Is this a new project?
├─ YES → Run /setup-project
│ ↓
│ Done! Use /protocol to start working
│
└─ NO → Do you need to update?
├─ YES → Run /update-project
└─ NO → Just use /protocol or /continue
Related Documentation
| Topic |
Document |
| Complete setup walkthrough |
USER-JOURNEY.md |
| Extension specifications |
EXTENSION-SPEC.md |
| Agent details |
AGENTS.md |
| Configuration options |
CONFIGURATION.md |
| Customization guide |
CUSTOMIZATION.md |
1---2name: decision-guide-when-to-use-what3description: A comprehensive guide to help you choose the right tools, commands, agents, and approaches in Claude Copilot.4---5# Decision Guide: When to Use What67A comprehensive guide to help you choose the right tools, commands, agents, and approaches in Claude Copilot.89*Note: This is the complete reference. For a quick overview integrated into Claude Code's instructions, see [CLAUDE.md](../CLAUDE.md#quick-decision-guide).*1011---1213## Feature Selection1415### Feature Comparison1617| Feature | Invocation | Persistence | Best For | When NOT to Use |18|---------|------------|-------------|----------|-----------------|19| **Memory** | Auto | Cross-session | Context preservation, decisions, lessons | Short-term notes, temporary data |20| **Agents** | Protocol | Session | Expert tasks, complex work | Simple commands, quick tasks |21| **Skills** | Auto | On-demand | Reusable patterns, workflows | One-off solutions |22| **Commands** | Manual | Session | Quick shortcuts, workflows | Complex multi-step processes |23| **Extensions** | Auto | Permanent | Team standards, custom methodologies | Personal preferences |2425### Memory vs Skills vs Extensions2627| Question | Use Memory | Use Skills | Use Extensions |28|----------|------------|------------|----------------|29| Does it change per project? | ✓ | | |30| Is it a one-time decision? | ✓ | | |31| Is it a reusable pattern? | | ✓ | |32| Does the whole team need it? | | | ✓ |33| Is it company-specific? | | | ✓ |34| Does it override base behavior? | | | ✓ |35| Is it a lesson learned? | ✓ | | |36| Is it a workflow/automation? | | ✓ | |3738---3940## Command Selection4142### Command Decision Matrix4344| I want to... | Command | When | Where to Run |45|--------------|---------|------|--------------|46| Set up Claude Copilot first time | `/setup` | Once per machine | `~/.claude/copilot` |47| Add Copilot to new project | `/setup-project` | Once per project | Project root |48| Update project files | `/update-project` | After framework updates | Project root |49| Update framework itself | `/update-copilot` | When new version available | Any directory |50| Create team knowledge | `/knowledge-copilot` | Once per team/company | Any directory |51| Start fresh work | `/protocol [task]` | Each work session | Project root |52| Resume previous work | `/continue [stream]` | When returning to work | Project root |53| Run parallel work streams | `/orchestrate start` | Set up worktrees, launch via Task tool | Project root |54| Monitor orchestration | `/orchestrate status` | During parallel execution | Project root |55| Merge completed streams | `/orchestrate merge` | After streams complete | Project root |56| Verify MCP servers | `/mcp` | After setup, troubleshooting | Project root |5758**Command Arguments:**59- `/protocol <task>` - Auto-detect task type and route to agent (e.g., `/protocol fix the login bug`)60- `/continue <stream>` - Resume specific parallel stream (e.g., `/continue Stream-B`)6162### Setup Command Flowchart6364```65Are you using Claude Copilot for the first time?66├─ YES → Run /setup in ~/.claude/copilot67│ ↓68│ Is this a new project?69│ ├─ YES → Run /setup-project in project70│ └─ NO → Run /update-project in existing project71│72└─ NO → Did you update Claude Copilot?73 ├─ YES → Run /update-project in all projects74 └─ NO → Just use /protocol or /continue to work75```7677---7879## Agent Selection8081### Agent Routing Matrix8283| Task Type | Primary Agent | Secondary Agent(s) | Why This Flow |84|-----------|---------------|-------------------|---------------|85| **Bug Fix** | `qa` | → `me` | QA reproduces, Engineer fixes |86| **New Feature** | `sd` | → `uxd` → `uid` | Service → UX → Implementation |87| **API Design** | `ta` | → `me` → `doc` | Architecture → Code → Docs |88| **Security Review** | `sec` | | Security expertise |89| **Performance Issue** | `ta` | → `me` | Design analysis → Implementation |90| **UI Component** | `uids` | → `uid` | Visual design → Code |91| **Documentation** | `doc` | | Technical writing |92| **Deployment** | `do` | | DevOps expertise |93| **Architecture Decision** | `ta` | | System design |94| **User Research** | `sd` | | Experience strategy |95| **Copy/Messaging** | `cw` | | Content writing |9697### Scenario-Based Agent Selection9899| Scenario | Start With | Reasoning |100|----------|------------|-----------|101| "Users can't login" | `/protocol` (DEFECT) → `@agent-qa` | Needs reproduction and diagnosis |102| "Add dark mode" | `/protocol` (EXPERIENCE) → `@agent-sd` | Experience change requires journey analysis |103| "Optimize database queries" | `/protocol` (FEATURE) → `@agent-ta` | Architecture-level optimization |104| "Deploy to production" | `/protocol` (DEVOPS) → `@agent-do` | Infrastructure task |105| "Security audit" | `/protocol` (SECURITY) → `@agent-sec` | Security expertise required |106| "Write API docs" | `/protocol` (DOCUMENTATION) → `@agent-doc` | Documentation specialist |107| "Refactor auth module" | `/protocol` (ARCHITECTURE) → `@agent-ta` | Design decision needed |108| "Fix button alignment" | `/protocol` (DEFECT) → `@agent-uid` | UI implementation fix |109110---111112## Extension vs Override vs Skills113114### Extension Type Decision Tree115116```117Do you need to customize agent behavior?118├─ NO → Use base agents as-is119│120└─ YES → Do you want to replace the entire agent?121 ├─ YES → Use OVERRIDE (.override.md)122 │ Example: Completely custom methodology123 │124 └─ NO → Do you want to add/enhance sections?125 ├─ YES → Use EXTENSION (.extension.md)126 │ Example: Add checklists, templates127 │128 └─ NO → Do you just need to inject skills?129 └─ Use SKILLS (.skills.json)130 Example: Company-specific tools131```132133### Extension Type Comparison134135| Goal | Extension Type | File Pattern | Scope | Difficulty |136|------|----------------|--------------|-------|------------|137| Add company checklist | `extension` | `agent.extension.md` | Section merge | Easy |138| Replace methodology | `override` | `agent.override.md` | Full replacement | Hard |139| Inject team tools | `skills` | `agent.skills.json` | Skill list only | Medium |140| Enhance templates | `extension` | `agent.extension.md` | Section merge | Easy |141| Custom process flow | `override` | `agent.override.md` | Full replacement | Hard |142| Add domain knowledge | `skills` | `agent.skills.json` | Skill list only | Medium |143144---145146## Work Session Decisions147148### Starting Work Decision Matrix149150| Situation | Use | Why |151|-----------|-----|-----|152| Brand new task | `/protocol` or `/protocol <task>` | Classify and route to expert |153| Continuing yesterday's work | `/continue` | Load context from memory |154| Resume specific parallel stream | `/continue <stream>` | Jump directly to stream work |155| Quick question | Just ask | No need for protocol |156| Exploring ideas | Just ask | Protocol is for execution |157| Complex multi-step task | `/protocol` | Agent expertise needed |158| Resume after interruption | `/continue` | Context restoration |159160### Protocol vs Direct Conversation161162| Use /protocol When... | Use Direct Conversation When... |163|----------------------|--------------------------------|164| Building a feature | Asking a question |165| Fixing a bug | Exploring ideas |166| Making architecture changes | Getting quick help |167| Need expert guidance | Simple task |168| Multi-step process | Single action |169| Want structured approach | Want flexibility |170171---172173## Knowledge Repository Decisions174175### When to Create Knowledge Repository176177| Indicator | Action |178|-----------|--------|179| Team > 1 person | Create knowledge repo |180| Company has style guide | Document in knowledge repo |181| Custom methodologies | Create agent extensions |182| Repeated explanations | Document once in knowledge |183| Onboarding takes days | Create knowledge repo |184| Inconsistent outputs | Define standards in knowledge |185186### Knowledge vs Skills vs Memory187188| Type of Information | Store In | Why |189|---------------------|----------|-----|190| Company values | Knowledge | Permanent, team-wide |191| Voice/tone guidelines | Knowledge | Permanent, team-wide |192| Design system | Knowledge | Permanent, team-wide |193| Reusable code patterns | Skills | On-demand, reusable |194| Project decisions | Memory | Project-specific, temporal |195| Lessons learned | Memory | Project-specific, evolving |196| API standards | Knowledge | Team standard |197| Deployment process | Skills | Reusable workflow |198| Why we chose X | Memory | Project context |199200---201202## Stream Management Decisions203204### When to Use Streams205206| Scenario | Use Streams? | Why |207|----------|--------------|-----|208| Single developer, sequential work | No | Standard task flow is simpler |209| Multiple parallel work areas | Yes | Prevent file conflicts |210| Large feature with independent components | Yes | Parallel development |211| Bug fix while feature in progress | Yes | Isolate changes |212| Foundation → parallel → integration workflow | Yes | Enforce dependencies |213214### Stream Pattern215216**Foundation Phase:**217- Stream-A: Core infrastructure that parallel work depends on218- Example: Schema changes, base types, shared utilities219220**Parallel Phase:**221- Stream-B, Stream-C, Stream-D: Independent work streams222- Each touches different files223- Can run simultaneously in separate Claude Code sessions224- Example: Stream-B (API endpoints), Stream-C (UI components), Stream-D (tests)225226**Integration Phase:**227- Stream-Z: Combines parallel streams228- Example: Documentation, final validation, release prep229230### Stream Tools Usage231232| Command | When to Use | Example |233|------|-------------|---------|234| `tc stream list --json` | View all streams in initiative | See progress across parallel work |235| `tc stream get <id> --json` | Get detailed stream info | Check Stream-B status before resuming |236| `git diff` | Before creating tasks | Ensure no file conflicts with other streams |237238### Stream Metadata in Tasks239240When creating tasks with streams, include:241242```typescript243metadata: {244 streamId: "Stream-B", // Auto-generated ID245 streamName: "command-updates", // Human-readable name246 streamPhase: "parallel", // foundation | parallel | integration247 files: [ // Files this task will modify248 "path/to/file1.ts",249 "path/to/file2.md"250 ],251 streamDependencies: ["Stream-A"] // Must complete before this stream252}253```254255---256257## Troubleshooting Decisions258259### When Something Goes Wrong260261| Problem | Check First | Then Try | Last Resort |262|---------|-------------|----------|-------------|263| Command not found | Machine setup complete? | `/setup` in `~/.claude/copilot` | Reinstall |264| MCP not connecting | `.mcp.json` paths absolute? | Rebuild servers | Check Node version |265| Agent not routing | Is task description clear? | Rephrase request | Use agent directly |266| Memory not persisting | `WORKSPACE_ID` set? | Check `.mcp.json` | Rebuild memory server |267| Knowledge not found | Symlink exists? | `/knowledge-copilot` | Manual link |268| Skills not loading | Skills server running? | `/mcp` to verify | Check logs |269270---271272## Quick Reference Tables273274### Installation Decision Matrix275276| You Are... | Steps Required | Commands to Run |277|------------|----------------|-----------------|278| New solo user | Clone → Setup → Project | `/setup` → `/setup-project` |279| New team member | Clone → Setup → Link Knowledge → Project | `/setup` → `/knowledge-copilot` → `/setup-project` |280| Existing user, new project | Project setup only | `/setup-project` |281| Updating framework | Update → Rebuild → Sync projects | `/update-copilot` → `/update-project` |282283### Daily Workflow Matrix284285| Beginning of Day | During Day | End of Day |286|------------------|------------|------------|287| `/continue` to resume | Work naturally with agents | `initiative_update` to save progress |288| Or `/protocol` for new task | Route complex work to specialists | Document decisions in memory |289| Check `/mcp` if needed | Use skills as needed | Note lessons learned |290291---292293## Best Practices Summary294295### Do This296297| Context | Best Practice |298|---------|--------------|299| Starting work | Use `/continue` to load context |300| Complex tasks | Use `/protocol` to engage experts |301| Team standards | Create knowledge repository |302| Reusable patterns | Save as skills |303| Project decisions | Store in memory with context |304| End of session | Update initiative with progress |305306### Don't Do This307308| Anti-Pattern | Instead Do |309|--------------|-----------|310| Hardcode paths in extensions | Use relative paths |311| Store secrets in knowledge | Use environment variables |312| Override agents unnecessarily | Use extensions first |313| Skip memory updates | Update at end of session |314| Create skills for one-offs | Just solve the problem |315| Ignore agent routing | Trust the protocol |316317---318319## Decision Flowchart: Complete Setup320321```322START: Do you have Claude Copilot installed?323│324├─ NO → Clone to ~/.claude/copilot325│ ↓326│ Run /setup in ~/.claude/copilot327│ ↓328│ Are you part of a team?329│ ├─ YES → Clone team knowledge repo330│ │ Run /knowledge-copilot to link331│ │ ↓332│ └─ NO → Do you want to create team knowledge?333│ ├─ YES → Run /knowledge-copilot334│ └─ NO → Skip knowledge setup335│ ↓336│ Run /setup-project in your project337│ ↓338│ Done! Use /protocol to start working339│340└─ YES → Is this a new project?341 ├─ YES → Run /setup-project342 │ ↓343 │ Done! Use /protocol to start working344 │345 └─ NO → Do you need to update?346 ├─ YES → Run /update-project347 └─ NO → Just use /protocol or /continue348```349350---351352## Related Documentation353354| Topic | Document |355|-------|----------|356| Complete setup walkthrough | [USER-JOURNEY.md](USER-JOURNEY.md) |357| Extension specifications | [EXTENSION-SPEC.md](EXTENSION-SPEC.md) |358| Agent details | [AGENTS.md](AGENTS.md) |359| Configuration options | [CONFIGURATION.md](CONFIGURATION.md) |360| Customization guide | [CUSTOMIZATION.md](CUSTOMIZATION.md) |