kiro-agents 1.3.0 → 1.5.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +56 -227
- package/build/npm/bin/cli.js +0 -2
- package/build/npm/dist/aliases.md +43 -0
- package/build/npm/dist/modes.md +135 -86
- package/build/npm/dist/protocols/agent-activation.mdx +1 -1
- package/build/npm/dist/protocols/mode-management.mdx +139 -0
- package/build/npm/dist/protocols/mode-switching.mdx +84 -84
- package/package.json +26 -3
package/README.md
CHANGED
|
@@ -1,266 +1,95 @@
|
|
|
1
1
|
# kiro-agents
|
|
2
2
|
|
|
3
|
-
|
|
4
|
-
|
|
5
|
-
## What is kiro-agents?
|
|
6
|
-
|
|
7
|
-
kiro-agents is a steering document system that extends Kiro IDE with:
|
|
8
|
-
|
|
9
|
-
- **Agent System** - Create and manage specialized AI agents with `/agent` and `/agents` commands
|
|
10
|
-
- **Mode System** - Switch between interaction styles with `/mode` command (vibe/spec modes)
|
|
11
|
-
- **Strict Mode** - Precision mode for experimental development with `/strict` command
|
|
12
|
-
- **Interactive Workflows** - Choice-based interactions that minimize cognitive load
|
|
13
|
-
- **Progressive Enhancement** - Start minimal, add agents as needed
|
|
14
|
-
|
|
15
|
-
## Installation
|
|
16
|
-
|
|
17
|
-
### Option 1: Global Installation (npm)
|
|
3
|
+
AI agent system for Kiro IDE. Create specialized agents, switch interaction modes, and reduce cognitive load during development.
|
|
18
4
|
|
|
19
5
|
```bash
|
|
20
6
|
npx kiro-agents
|
|
21
|
-
# or
|
|
22
|
-
bunx kiro-agents
|
|
23
|
-
```
|
|
24
|
-
|
|
25
|
-
This installs steering documents to `~/.kiro/steering/` (global, available in all workspaces).
|
|
26
|
-
|
|
27
|
-
### Option 2: Workspace Installation (Kiro Power)
|
|
28
|
-
|
|
29
|
-
1. Open Kiro IDE
|
|
30
|
-
2. Open Powers panel (Ghosty icon ⚡)
|
|
31
|
-
3. Click "Add power from GitHub"
|
|
32
|
-
4. Enter: `https://github.com/theadd/kiro-agents`
|
|
33
|
-
5. Click Install
|
|
34
|
-
|
|
35
|
-
This installs to `.kiro/powers/kiro-agents/` (workspace-specific).
|
|
36
|
-
|
|
37
|
-
**Differences:**
|
|
38
|
-
- **npm**: Global installation, available in all workspaces
|
|
39
|
-
- **Power**: Workspace-specific, auto-updates, better Kiro integration
|
|
40
|
-
|
|
41
|
-
## Core Features
|
|
42
|
-
|
|
43
|
-
### Agent System
|
|
44
|
-
|
|
45
|
-
Create specialized AI agents for different tasks. Each agent has:
|
|
46
|
-
|
|
47
|
-
- Defined capabilities and responsibilities
|
|
48
|
-
- Custom interaction protocols
|
|
49
|
-
- Specific workflows and tools
|
|
50
|
-
- Integration with other agents
|
|
51
|
-
|
|
52
|
-
**Commands:**
|
|
53
|
-
- `/agents` - Interactive agent management (visual menu)
|
|
54
|
-
- `/agents {name}` - Activate specific agent directly
|
|
55
|
-
|
|
56
|
-
**Example workflow:**
|
|
57
|
-
```
|
|
58
|
-
User: /agents
|
|
59
|
-
AI: [Shows available agents with numbered choices]
|
|
60
|
-
User: 2 (Create new agent)
|
|
61
|
-
AI: [Guides through agent creation wizard]
|
|
62
7
|
```
|
|
63
8
|
|
|
64
|
-
|
|
65
|
-
- Created interactively through wizard
|
|
66
|
-
- Modified and updated
|
|
67
|
-
- Activated on-demand
|
|
68
|
-
- Coordinated with other agents
|
|
69
|
-
|
|
70
|
-
### Mode System
|
|
71
|
-
|
|
72
|
-
Switch between interaction styles based on your needs:
|
|
73
|
-
|
|
74
|
-
**Vibe Mode** - Flexible, conversational development
|
|
75
|
-
- Quick iterations and prototyping
|
|
76
|
-
- Direct code changes
|
|
77
|
-
- Exploratory development
|
|
78
|
-
- Minimal ceremony
|
|
79
|
-
|
|
80
|
-
**Spec Mode** - Structured feature development
|
|
81
|
-
- Formal requirements (EARS/INCOSE compliant)
|
|
82
|
-
- Design with correctness properties
|
|
83
|
-
- Task-based implementation
|
|
84
|
-
- Property-based testing focus
|
|
85
|
-
|
|
86
|
-
**Commands:**
|
|
87
|
-
- `/modes` - Interactive mode management (visual menu)
|
|
88
|
-
- `/modes vibe` - Switch to flexible development
|
|
89
|
-
- `/modes spec` - Switch to structured planning
|
|
90
|
-
|
|
91
|
-
### Strict Mode
|
|
92
|
-
|
|
93
|
-
Precision mode for experimental or critical development:
|
|
94
|
-
|
|
95
|
-
- Blocks execution on ambiguous input
|
|
96
|
-
- Requires explicit clarification
|
|
97
|
-
- Prevents assumption propagation
|
|
98
|
-
- Ideal for architectural decisions
|
|
99
|
-
|
|
100
|
-
**Commands:**
|
|
101
|
-
- `/strict` - Interactive control (visual buttons)
|
|
102
|
-
- `/strict on` - Activate precision mode
|
|
103
|
-
- `/strict off` - Return to normal mode
|
|
104
|
-
|
|
105
|
-
### Interactive Workflows
|
|
106
|
-
|
|
107
|
-
All interactions use choice-based patterns:
|
|
9
|
+
## Features
|
|
108
10
|
|
|
109
|
-
-
|
|
110
|
-
-
|
|
111
|
-
-
|
|
112
|
-
-
|
|
11
|
+
- **Agents** - Specialized AI personas with defined capabilities (`/agents`)
|
|
12
|
+
- **Modes** - Switch between vibe (flexible) and spec (structured) workflows (`/modes`)
|
|
13
|
+
- **Strict Mode** - Precision mode that blocks on ambiguity (`/strict`)
|
|
14
|
+
- **Zero Config** - Works immediately after installation ⇨ <kbd>/agents</kbd>
|
|
113
15
|
|
|
114
|
-
|
|
115
|
-
|
|
116
|
-
## How It Works
|
|
117
|
-
|
|
118
|
-
kiro-agents provides steering documents that Kiro loads to enhance AI capabilities:
|
|
119
|
-
|
|
120
|
-
**npm installation:**
|
|
121
|
-
- Installs to `~/.kiro/steering/` (global)
|
|
122
|
-
- Available in all workspaces
|
|
123
|
-
- Manual updates via re-running npx/bunx
|
|
124
|
-
|
|
125
|
-
**Power installation:**
|
|
126
|
-
- Installs to `.kiro/powers/kiro-agents/` (workspace)
|
|
127
|
-
- Auto-updates from GitHub
|
|
128
|
-
- Better integration with Kiro ecosystem
|
|
129
|
-
- Keyword-based activation
|
|
130
|
-
|
|
131
|
-
Both methods provide:
|
|
132
|
-
1. New slash commands (`/agents`, `/modes`, `/strict`)
|
|
133
|
-
2. Protocols for AI behavior
|
|
134
|
-
3. Agent and mode systems
|
|
135
|
-
4. Optimized interactions for reduced cognitive load
|
|
136
|
-
|
|
137
|
-
The system adds minimal overhead while enabling progressive enhancement.
|
|
138
|
-
|
|
139
|
-
## File Structure After Installation
|
|
140
|
-
|
|
141
|
-
**npm installation** (`~/.kiro/steering/`):
|
|
142
|
-
```
|
|
143
|
-
~/.kiro/steering/
|
|
144
|
-
├── agent-system.md
|
|
145
|
-
├── modes-system.md
|
|
146
|
-
└── agent-system/
|
|
147
|
-
├── strict-mode.md
|
|
148
|
-
├── kiro-spec-mode.md
|
|
149
|
-
├── kiro-vibe-mode.md
|
|
150
|
-
├── interactions/
|
|
151
|
-
└── tools/
|
|
152
|
-
```
|
|
16
|
+
## Installation
|
|
153
17
|
|
|
154
|
-
**
|
|
155
|
-
```
|
|
156
|
-
|
|
157
|
-
├── POWER.md # Power metadata
|
|
158
|
-
├── mcp.json # MCP server config (empty for now)
|
|
159
|
-
└── steering/
|
|
160
|
-
├── agent-system.md # Core system (always loaded)
|
|
161
|
-
├── agents.md # Interactive menu (manual)
|
|
162
|
-
├── modes-system.md # Core system (always loaded)
|
|
163
|
-
├── modes.md # Interactive menu (manual)
|
|
164
|
-
├── strict-mode.md # Core system (always loaded)
|
|
165
|
-
├── strict.md # Interactive control (manual)
|
|
166
|
-
├── interactions/ # Interaction patterns
|
|
167
|
-
└── modes/ # Mode definitions
|
|
18
|
+
**Global (npm):**
|
|
19
|
+
```bash
|
|
20
|
+
npx kiro-agents
|
|
168
21
|
```
|
|
22
|
+
Installs to `~/.kiro/steering/kiro-agents/` - available in all workspaces.
|
|
169
23
|
|
|
170
|
-
|
|
24
|
+
<!--
|
|
25
|
+
**Workspace (Kiro Power):**
|
|
26
|
+
1. Open Powers panel in Kiro IDE
|
|
27
|
+
2. Add from GitHub: `https://github.com/theadd/kiro-agents`
|
|
171
28
|
|
|
172
|
-
|
|
29
|
+
Installs to `.kiro/powers/kiro-agents/` - auto-updates, workspace-specific.
|
|
30
|
+
-->
|
|
173
31
|
|
|
174
|
-
|
|
32
|
+
## Quick Start
|
|
175
33
|
|
|
34
|
+
**Create an agent:**
|
|
176
35
|
```
|
|
177
|
-
|
|
178
|
-
|
|
179
|
-
|
|
180
|
-
AI: What type of agent? [Shows options]
|
|
181
|
-
User: 1 (Code-focused agent)
|
|
182
|
-
AI: [Collects name, capabilities, workflows]
|
|
183
|
-
AI: Agent created! Activate now? [Yes/No]
|
|
36
|
+
/agents
|
|
37
|
+
→ Choose "Create new agent"
|
|
38
|
+
→ Follow wizard
|
|
184
39
|
```
|
|
185
40
|
|
|
186
|
-
|
|
187
|
-
|
|
188
|
-
```
|
|
189
|
-
User: /modes spec
|
|
190
|
-
AI: [Loads spec mode protocols]
|
|
191
|
-
AI: What feature do you want to work on?
|
|
192
|
-
User: User authentication system
|
|
193
|
-
AI: [Guides through requirements → design → tasks]
|
|
194
|
-
```
|
|
195
|
-
|
|
196
|
-
### Using Strict Mode
|
|
197
|
-
|
|
41
|
+
**Switch modes:**
|
|
198
42
|
```
|
|
199
|
-
|
|
200
|
-
|
|
201
|
-
User: Add caching to the API
|
|
202
|
-
AI: [Asks clarifying questions about cache strategy, TTL, invalidation, etc.]
|
|
43
|
+
/modes spec # Structured feature development
|
|
44
|
+
/modes vibe # Flexible, conversational coding
|
|
203
45
|
```
|
|
204
46
|
|
|
205
|
-
|
|
206
|
-
|
|
47
|
+
**Enable precision mode:**
|
|
207
48
|
```
|
|
208
|
-
|
|
209
|
-
[Work on feature planning]
|
|
210
|
-
User: /agents kiro-master
|
|
211
|
-
[Use kiro-master capabilities while in spec mode]
|
|
49
|
+
/strict on # Blocks on ambiguous input
|
|
212
50
|
```
|
|
213
51
|
|
|
214
|
-
##
|
|
52
|
+
## How It Works
|
|
215
53
|
|
|
216
|
-
|
|
217
|
-
- Core system adds small context footprint
|
|
218
|
-
- Agents created progressively as needed
|
|
219
|
-
- Only load what you use
|
|
54
|
+
kiro-agents installs steering documents (markdown files with AI instructions) that extend Kiro's capabilities:
|
|
220
55
|
|
|
221
|
-
**
|
|
222
|
-
-
|
|
223
|
-
-
|
|
224
|
-
- Single focus per message
|
|
225
|
-
- Clear context maintenance
|
|
56
|
+
- **Agents** - Stored in `.kiro/agents/{name}.md`, activated via `/agents {name}`
|
|
57
|
+
- **Modes** - Define interaction style (vibe = flexible, spec = structured)
|
|
58
|
+
- **Strict Mode** - Requires explicit clarification before execution
|
|
226
59
|
|
|
227
|
-
|
|
228
|
-
- Switch between modes as needs change
|
|
229
|
-
- Combine modes with agents
|
|
230
|
-
- Adapt to your working style
|
|
60
|
+
All interactions use numbered choices to reduce cognitive load.
|
|
231
61
|
|
|
232
|
-
|
|
233
|
-
- Create custom agents for your needs
|
|
234
|
-
- Define specialized workflows
|
|
235
|
-
- Integrate with existing Kiro features
|
|
62
|
+
## Use Cases
|
|
236
63
|
|
|
237
|
-
|
|
64
|
+
**Specialized Development:**
|
|
65
|
+
- Create agents for frontend, backend, testing, DevOps
|
|
66
|
+
- Each agent has domain-specific knowledge and workflows
|
|
67
|
+
- Switch agents based on current task
|
|
238
68
|
|
|
239
|
-
|
|
69
|
+
**Workflow Flexibility:**
|
|
70
|
+
- Vibe mode for prototyping and exploration
|
|
71
|
+
- Spec mode for production features with formal requirements
|
|
72
|
+
- Combine modes with agents for maximum control
|
|
240
73
|
|
|
241
|
-
|
|
242
|
-
-
|
|
243
|
-
-
|
|
244
|
-
-
|
|
245
|
-
- Collaborative workflows
|
|
74
|
+
**Precision Work:**
|
|
75
|
+
- Enable strict mode for architectural decisions
|
|
76
|
+
- Prevents AI from making assumptions
|
|
77
|
+
- Forces explicit clarification on ambiguous requests
|
|
246
78
|
|
|
247
|
-
|
|
79
|
+
## Why kiro-agents?
|
|
248
80
|
|
|
249
|
-
|
|
250
|
-
-
|
|
251
|
-
-
|
|
252
|
-
-
|
|
81
|
+
Traditional AI assistants overwhelm you with context and options. kiro-agents:
|
|
82
|
+
- Starts minimal, grows with your needs
|
|
83
|
+
- Uses numbered choices instead of open-ended questions
|
|
84
|
+
- Maintains single focus per interaction
|
|
85
|
+
- Reduces decision fatigue during long sessions
|
|
253
86
|
|
|
254
|
-
##
|
|
87
|
+
## Links
|
|
255
88
|
|
|
256
|
-
|
|
89
|
+
- [npm Package](https://www.npmjs.com/package/kiro-agents)
|
|
90
|
+
- [GitHub](https://github.com/Theadd/kiro-agents)
|
|
91
|
+
- [Kiro IDE](https://kiro.dev)
|
|
257
92
|
|
|
258
93
|
## License
|
|
259
94
|
|
|
260
95
|
MIT
|
|
261
|
-
|
|
262
|
-
## Links
|
|
263
|
-
|
|
264
|
-
- [GitHub Repository](https://github.com/theadd/kiro-agents)
|
|
265
|
-
- [npm Package](https://www.npmjs.com/package/kiro-agents)
|
|
266
|
-
- [Kiro IDE](https://kiro.dev)
|
package/build/npm/bin/cli.js
CHANGED
|
@@ -27,8 +27,6 @@ var __filename2 = fileURLToPath(import.meta.url);
|
|
|
27
27
|
var __dirname2 = dirname(__filename2);
|
|
28
28
|
var INSTALL_DIR = join(homedir(), ".kiro", "steering", "kiro-agents");
|
|
29
29
|
var FILES_TO_INSTALL = [
|
|
30
|
-
"agent-system.md",
|
|
31
|
-
"modes-system.md",
|
|
32
30
|
"strict-mode.md",
|
|
33
31
|
"agents.md",
|
|
34
32
|
"modes.md",
|
|
@@ -75,3 +75,46 @@ spec
|
|
|
75
75
|
````
|
|
76
76
|
|
|
77
77
|
User types: `/spec` → AI responds: `spec`
|
|
78
|
+
|
|
79
|
+
## Agent System Alias
|
|
80
|
+
|
|
81
|
+
The agent activation command uses parameter substitution to load agent definitions dynamically:
|
|
82
|
+
|
|
83
|
+
<alias>
|
|
84
|
+
<trigger>/agents {agent_name}</trigger>
|
|
85
|
+
<definition>
|
|
86
|
+
## Agent Activation: {agent_name}
|
|
87
|
+
|
|
88
|
+
You are now activating the **{agent_name}** agent.
|
|
89
|
+
|
|
90
|
+
**Load and execute activation protocol:**
|
|
91
|
+
1. Read `.kiro/agents/{agent_name}.md` into context
|
|
92
|
+
2. Read `~/.kiro/steering/kiro-agents/protocols/agent-activation.mdx` into context
|
|
93
|
+
3. Follow all steps from the "Agent Activation Steps" section in agent-activation.mdx
|
|
94
|
+
4. Use `{agent_name}` as the agent identifier throughout the protocol
|
|
95
|
+
</definition>
|
|
96
|
+
</alias>
|
|
97
|
+
|
|
98
|
+
This alias enables users to activate any agent with `/agents {name}` syntax.
|
|
99
|
+
|
|
100
|
+
|
|
101
|
+
## Mode System Alias
|
|
102
|
+
|
|
103
|
+
The mode switching command uses parameter substitution to load mode definitions dynamically:
|
|
104
|
+
|
|
105
|
+
<alias>
|
|
106
|
+
<trigger>/modes {mode_name}</trigger>
|
|
107
|
+
<definition>
|
|
108
|
+
## Mode Switch: {mode_name}
|
|
109
|
+
|
|
110
|
+
You are now switching to **{mode_name} mode**.
|
|
111
|
+
|
|
112
|
+
**Load and execute mode switching protocol:**
|
|
113
|
+
1. Read `kiro-{mode_name}-mode.md` from agent-system directory into context
|
|
114
|
+
2. Read `~/.kiro/steering/kiro-agents/protocols/mode-switching.mdx` into context
|
|
115
|
+
3. Follow all steps from the "Mode Switch Steps" section in mode-switching.mdx
|
|
116
|
+
4. Use `{mode_name}` as the mode identifier throughout the protocol
|
|
117
|
+
</definition>
|
|
118
|
+
</alias>
|
|
119
|
+
|
|
120
|
+
This alias enables users to switch modes with `/modes {name}` syntax.
|
package/build/npm/dist/modes.md
CHANGED
|
@@ -8,93 +8,142 @@ keywords: ["modes", "vibe", "spec", "workflow", "interactive"]
|
|
|
8
8
|
|
|
9
9
|
You are now in **mode management** using chit-chat interaction protocol.
|
|
10
10
|
|
|
11
|
-
##
|
|
11
|
+
## Mode Management Steps
|
|
12
|
+
|
|
13
|
+
When entering mode management:
|
|
14
|
+
|
|
15
|
+
### Step 1: Load Chit-Chat Mode
|
|
16
|
+
|
|
17
|
+
Apply protocols from `chit-chat.md` steering document:
|
|
18
|
+
- Use diff blocks to show progress
|
|
19
|
+
- Provide numbered choice lists (4-6 options)
|
|
20
|
+
- Maintain single focus per message
|
|
21
|
+
- Use visual formatting (bold, code blocks, lists)
|
|
22
|
+
|
|
23
|
+
Begin with a diff block showing:
|
|
24
|
+
```diff
|
|
25
|
+
👉 Mode management
|
|
26
|
+
⏳ Mode selection
|
|
27
|
+
```
|
|
28
|
+
|
|
29
|
+
### Step 2: Detect Available Modes
|
|
30
|
+
|
|
31
|
+
Scan agent-system directory for `kiro-*-mode.md` files:
|
|
32
|
+
- Extract mode names from filenames (remove `kiro-` prefix and `-mode.md` suffix)
|
|
33
|
+
- Read frontmatter for descriptions
|
|
34
|
+
- Identify current mode from context (if any)
|
|
35
|
+
|
|
36
|
+
**Available Kiro modes:**
|
|
37
|
+
- **vibe** - Flexible, conversational development assistance
|
|
38
|
+
- **spec** - Structured feature development with requirements, design, and tasks
|
|
39
|
+
|
|
40
|
+
### Step 3: Present Mode Selection
|
|
41
|
+
|
|
42
|
+
Use this response structure:
|
|
43
|
+
|
|
44
|
+
```diff
|
|
45
|
+
👉 Mode management
|
|
46
|
+
⏳ Mode selection
|
|
47
|
+
```
|
|
48
|
+
|
|
49
|
+
**Current mode:** [current_mode or "none"]
|
|
50
|
+
|
|
51
|
+
**Available Kiro modes:**
|
|
52
|
+
|
|
53
|
+
- **vibe** - Flexible, conversational development assistance
|
|
54
|
+
- **spec** - Structured feature development with requirements, design, and tasks
|
|
55
|
+
|
|
56
|
+
**What would you like to do?**
|
|
57
|
+
|
|
58
|
+
1. **Switch to vibe mode** - Flexible development and quick iterations
|
|
59
|
+
2. **Switch to spec mode** - Structured feature planning and implementation
|
|
60
|
+
3. **View mode details** - See full mode capabilities and protocols
|
|
61
|
+
4. **Mode comparison** - Understand differences between modes
|
|
62
|
+
5. **Stay in current mode** - Continue with current mode
|
|
63
|
+
6. **Help** - Learn about mode system
|
|
64
|
+
|
|
65
|
+
### Step 4: Handle User Choice
|
|
66
|
+
|
|
67
|
+
Based on user selection:
|
|
68
|
+
|
|
69
|
+
#### Option 1 - Switch to vibe
|
|
70
|
+
|
|
71
|
+
- Execute `/modes vibe` command
|
|
72
|
+
- Load vibe mode protocols
|
|
73
|
+
- Begin flexible interaction
|
|
74
|
+
|
|
75
|
+
#### Option 2 - Switch to spec
|
|
76
|
+
|
|
77
|
+
- Execute `/modes spec` command
|
|
78
|
+
- Load spec mode protocols
|
|
79
|
+
- Begin structured workflow
|
|
80
|
+
|
|
81
|
+
#### Option 3 - View mode details
|
|
82
|
+
|
|
83
|
+
- Show numbered list of modes
|
|
84
|
+
- User selects mode to view
|
|
85
|
+
- Display full mode definition
|
|
86
|
+
- Explain capabilities and use cases
|
|
87
|
+
- Offer to switch to that mode
|
|
88
|
+
|
|
89
|
+
#### Option 4 - Mode comparison
|
|
90
|
+
|
|
91
|
+
- Show side-by-side comparison:
|
|
92
|
+
- **Vibe Mode**: Flexible, conversational, quick iterations, no formal workflow
|
|
93
|
+
- **Spec Mode**: Structured workflow with requirements → design → tasks, approval gates
|
|
94
|
+
- Explain when to use each mode:
|
|
95
|
+
- Use vibe for: Quick fixes, exploration, prototyping, iterative development
|
|
96
|
+
- Use spec for: Complex features, team collaboration, formal planning, documentation
|
|
97
|
+
- Highlight key differences:
|
|
98
|
+
- Workflow: None vs. Structured phases
|
|
99
|
+
- Approval: Direct changes vs. Approval gates
|
|
100
|
+
- Documentation: Minimal vs. Comprehensive
|
|
101
|
+
- Help user choose appropriate mode
|
|
102
|
+
|
|
103
|
+
#### Option 5 - Stay in current mode
|
|
104
|
+
|
|
105
|
+
- Confirm staying in current mode
|
|
106
|
+
- Return to normal interaction
|
|
107
|
+
- Preserve current state
|
|
108
|
+
|
|
109
|
+
#### Option 6 - Help
|
|
110
|
+
|
|
111
|
+
Explain mode system:
|
|
112
|
+
- **What are modes?** - Different interaction styles for different workflows
|
|
113
|
+
- **How to switch?** - Use `/modes {name}` or `/modes` for interactive menu
|
|
114
|
+
- **Mode benefits:**
|
|
115
|
+
- Vibe: Fast iteration, flexible approach
|
|
116
|
+
- Spec: Structured planning, comprehensive documentation
|
|
117
|
+
- **Mode coordination:**
|
|
118
|
+
- Modes can be combined with agents
|
|
119
|
+
- File changes preserved when switching
|
|
120
|
+
- Workflow state resets when switching
|
|
121
|
+
- **Usage examples:**
|
|
122
|
+
- Quick bug fix → Use vibe mode
|
|
123
|
+
- New feature with requirements → Use spec mode
|
|
124
|
+
- Refactoring existing code → Use vibe mode
|
|
125
|
+
- Team feature with documentation → Use spec mode
|
|
126
|
+
|
|
127
|
+
### Step 5: Maintain Chit-Chat Mode
|
|
128
|
+
|
|
129
|
+
Continue using diff blocks and numbered choices throughout mode management session.
|
|
130
|
+
|
|
131
|
+
After each action:
|
|
132
|
+
- Update diff block with progress
|
|
133
|
+
- Show current focus
|
|
134
|
+
- Provide next action choices
|
|
135
|
+
- Allow going back or canceling
|
|
136
|
+
|
|
137
|
+
**Context Preservation:**
|
|
138
|
+
- Use diff blocks to show progress
|
|
139
|
+
- Preserve user decisions
|
|
140
|
+
- Allow going back to previous step
|
|
141
|
+
- Enable canceling operations
|
|
142
|
+
|
|
143
|
+
---
|
|
144
|
+
|
|
145
|
+
**Mode management active. Present choices to user.**
|
|
12
146
|
|
|
13
|
-
Apply protocols from `chit-chat.md` steering document:
|
|
14
|
-
- Use diff blocks to show progress
|
|
15
|
-
- Provide numbered choice lists (4-6 options)
|
|
16
|
-
- Maintain single focus per message
|
|
17
|
-
- Use visual formatting (bold, code blocks, lists)
|
|
18
|
-
|
|
19
|
-
## Step 2: Detect Available Modes
|
|
20
|
-
|
|
21
|
-
Scan agent-system directory for `kiro-*-mode.md` files:
|
|
22
|
-
- Extract mode names from filenames
|
|
23
|
-
- Read frontmatter for descriptions
|
|
24
|
-
- Identify current mode from context
|
|
25
|
-
|
|
26
|
-
## Step 3: Present Mode Selection
|
|
27
|
-
|
|
28
|
-
Use this response structure:
|
|
29
|
-
|
|
30
|
-
```diff
|
|
31
|
-
👉 Mode management
|
|
32
|
-
⏳ Mode selection
|
|
33
|
-
```
|
|
34
|
-
|
|
35
|
-
**Current mode:** [current_mode]
|
|
36
|
-
|
|
37
|
-
**Available Kiro modes:**
|
|
38
|
-
|
|
39
|
-
- **vibe** - Flexible, conversational development assistance
|
|
40
|
-
- **spec** - Structured feature development with requirements, design, and tasks
|
|
41
|
-
|
|
42
|
-
**What would you like to do?**
|
|
43
|
-
|
|
44
|
-
1. **Switch to vibe mode** - Flexible development and quick iterations
|
|
45
|
-
2. **Switch to spec mode** - Structured feature planning and implementation
|
|
46
|
-
3. **View mode details** - See full mode capabilities and protocols
|
|
47
|
-
4. **Mode comparison** - Understand differences between modes
|
|
48
|
-
5. **Stay in current mode** - Continue with current mode
|
|
49
|
-
6. **Help** - Learn about mode system
|
|
50
|
-
|
|
51
|
-
## Step 4: Handle User Choice
|
|
52
|
-
|
|
53
|
-
Based on user selection:
|
|
54
|
-
|
|
55
|
-
**Option 1 - Switch to vibe:**
|
|
56
|
-
- Execute `/modes vibe` command
|
|
57
|
-
- Load vibe mode protocols
|
|
58
|
-
- Begin flexible interaction
|
|
59
|
-
|
|
60
|
-
**Option 2 - Switch to spec:**
|
|
61
|
-
- Execute `/modes spec` command
|
|
62
|
-
- Load spec mode protocols
|
|
63
|
-
- Begin structured workflow
|
|
64
|
-
|
|
65
|
-
**Option 3 - View mode details:**
|
|
66
|
-
- Show numbered list of modes
|
|
67
|
-
- User selects mode to view
|
|
68
|
-
- Display full mode definition
|
|
69
|
-
- Explain capabilities and use cases
|
|
70
|
-
- Offer to switch to that mode
|
|
71
|
-
|
|
72
|
-
**Option 4 - Mode comparison:**
|
|
73
|
-
- Show side-by-side comparison
|
|
74
|
-
- Explain when to use each mode
|
|
75
|
-
- Highlight key differences
|
|
76
|
-
- Help user choose appropriate mode
|
|
77
|
-
|
|
78
|
-
**Option 5 - Stay in current mode:**
|
|
79
|
-
- Confirm staying in current mode
|
|
80
|
-
- Return to normal interaction
|
|
81
|
-
- Preserve current state
|
|
82
|
-
|
|
83
|
-
**Option 6 - Help:**
|
|
84
|
-
- Explain mode system
|
|
85
|
-
- Show switching commands
|
|
86
|
-
- Describe mode benefits
|
|
87
|
-
- Provide usage examples
|
|
88
|
-
|
|
89
|
-
## Step 5: Maintain Chit-Chat Mode
|
|
90
|
-
|
|
91
|
-
Continue using diff blocks and numbered choices throughout mode management.
|
|
92
|
-
|
|
93
|
-
After each action:
|
|
94
|
-
- Update diff block with progress
|
|
95
|
-
- Show current focus
|
|
96
|
-
- Provide next action choices
|
|
97
|
-
- Allow going back or canceling
|
|
98
147
|
|
|
99
148
|
---
|
|
100
149
|
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
# Agent Activation Protocol
|
|
2
2
|
|
|
3
|
-
This file contains the detailed instructions for activating an agent. It is referenced by the `/agents {agent_name}` alias in `
|
|
3
|
+
This file contains the detailed instructions for activating an agent. It is referenced by the `/agents {agent_name}` alias in `aliases.md`.
|
|
4
4
|
|
|
5
5
|
## Agent Activation Steps
|
|
6
6
|
|
|
@@ -0,0 +1,139 @@
|
|
|
1
|
+
# Mode Management Protocol
|
|
2
|
+
|
|
3
|
+
This file contains the detailed instructions for interactive mode management. It is referenced when `/modes` is executed without parameters.
|
|
4
|
+
|
|
5
|
+
## Mode Management Steps
|
|
6
|
+
|
|
7
|
+
When entering mode management:
|
|
8
|
+
|
|
9
|
+
### Step 1: Load Chit-Chat Mode
|
|
10
|
+
|
|
11
|
+
Apply protocols from `chit-chat.md` steering document:
|
|
12
|
+
- Use diff blocks to show progress
|
|
13
|
+
- Provide numbered choice lists (4-6 options)
|
|
14
|
+
- Maintain single focus per message
|
|
15
|
+
- Use visual formatting (bold, code blocks, lists)
|
|
16
|
+
|
|
17
|
+
Begin with a diff block showing:
|
|
18
|
+
```diff
|
|
19
|
+
👉 Mode management
|
|
20
|
+
⏳ Mode selection
|
|
21
|
+
```
|
|
22
|
+
|
|
23
|
+
### Step 2: Detect Available Modes
|
|
24
|
+
|
|
25
|
+
Scan agent-system directory for `kiro-*-mode.md` files:
|
|
26
|
+
- Extract mode names from filenames (remove `kiro-` prefix and `-mode.md` suffix)
|
|
27
|
+
- Read frontmatter for descriptions
|
|
28
|
+
- Identify current mode from context (if any)
|
|
29
|
+
|
|
30
|
+
**Available Kiro modes:**
|
|
31
|
+
- **vibe** - Flexible, conversational development assistance
|
|
32
|
+
- **spec** - Structured feature development with requirements, design, and tasks
|
|
33
|
+
|
|
34
|
+
### Step 3: Present Mode Selection
|
|
35
|
+
|
|
36
|
+
Use this response structure:
|
|
37
|
+
|
|
38
|
+
```diff
|
|
39
|
+
👉 Mode management
|
|
40
|
+
⏳ Mode selection
|
|
41
|
+
```
|
|
42
|
+
|
|
43
|
+
**Current mode:** [current_mode or "none"]
|
|
44
|
+
|
|
45
|
+
**Available Kiro modes:**
|
|
46
|
+
|
|
47
|
+
- **vibe** - Flexible, conversational development assistance
|
|
48
|
+
- **spec** - Structured feature development with requirements, design, and tasks
|
|
49
|
+
|
|
50
|
+
**What would you like to do?**
|
|
51
|
+
|
|
52
|
+
1. **Switch to vibe mode** - Flexible development and quick iterations
|
|
53
|
+
2. **Switch to spec mode** - Structured feature planning and implementation
|
|
54
|
+
3. **View mode details** - See full mode capabilities and protocols
|
|
55
|
+
4. **Mode comparison** - Understand differences between modes
|
|
56
|
+
5. **Stay in current mode** - Continue with current mode
|
|
57
|
+
6. **Help** - Learn about mode system
|
|
58
|
+
|
|
59
|
+
### Step 4: Handle User Choice
|
|
60
|
+
|
|
61
|
+
Based on user selection:
|
|
62
|
+
|
|
63
|
+
#### Option 1 - Switch to vibe
|
|
64
|
+
|
|
65
|
+
- Execute `/modes vibe` command
|
|
66
|
+
- Load vibe mode protocols
|
|
67
|
+
- Begin flexible interaction
|
|
68
|
+
|
|
69
|
+
#### Option 2 - Switch to spec
|
|
70
|
+
|
|
71
|
+
- Execute `/modes spec` command
|
|
72
|
+
- Load spec mode protocols
|
|
73
|
+
- Begin structured workflow
|
|
74
|
+
|
|
75
|
+
#### Option 3 - View mode details
|
|
76
|
+
|
|
77
|
+
- Show numbered list of modes
|
|
78
|
+
- User selects mode to view
|
|
79
|
+
- Display full mode definition
|
|
80
|
+
- Explain capabilities and use cases
|
|
81
|
+
- Offer to switch to that mode
|
|
82
|
+
|
|
83
|
+
#### Option 4 - Mode comparison
|
|
84
|
+
|
|
85
|
+
- Show side-by-side comparison:
|
|
86
|
+
- **Vibe Mode**: Flexible, conversational, quick iterations, no formal workflow
|
|
87
|
+
- **Spec Mode**: Structured workflow with requirements → design → tasks, approval gates
|
|
88
|
+
- Explain when to use each mode:
|
|
89
|
+
- Use vibe for: Quick fixes, exploration, prototyping, iterative development
|
|
90
|
+
- Use spec for: Complex features, team collaboration, formal planning, documentation
|
|
91
|
+
- Highlight key differences:
|
|
92
|
+
- Workflow: None vs. Structured phases
|
|
93
|
+
- Approval: Direct changes vs. Approval gates
|
|
94
|
+
- Documentation: Minimal vs. Comprehensive
|
|
95
|
+
- Help user choose appropriate mode
|
|
96
|
+
|
|
97
|
+
#### Option 5 - Stay in current mode
|
|
98
|
+
|
|
99
|
+
- Confirm staying in current mode
|
|
100
|
+
- Return to normal interaction
|
|
101
|
+
- Preserve current state
|
|
102
|
+
|
|
103
|
+
#### Option 6 - Help
|
|
104
|
+
|
|
105
|
+
Explain mode system:
|
|
106
|
+
- **What are modes?** - Different interaction styles for different workflows
|
|
107
|
+
- **How to switch?** - Use `/modes {name}` or `/modes` for interactive menu
|
|
108
|
+
- **Mode benefits:**
|
|
109
|
+
- Vibe: Fast iteration, flexible approach
|
|
110
|
+
- Spec: Structured planning, comprehensive documentation
|
|
111
|
+
- **Mode coordination:**
|
|
112
|
+
- Modes can be combined with agents
|
|
113
|
+
- File changes preserved when switching
|
|
114
|
+
- Workflow state resets when switching
|
|
115
|
+
- **Usage examples:**
|
|
116
|
+
- Quick bug fix → Use vibe mode
|
|
117
|
+
- New feature with requirements → Use spec mode
|
|
118
|
+
- Refactoring existing code → Use vibe mode
|
|
119
|
+
- Team feature with documentation → Use spec mode
|
|
120
|
+
|
|
121
|
+
### Step 5: Maintain Chit-Chat Mode
|
|
122
|
+
|
|
123
|
+
Continue using diff blocks and numbered choices throughout mode management session.
|
|
124
|
+
|
|
125
|
+
After each action:
|
|
126
|
+
- Update diff block with progress
|
|
127
|
+
- Show current focus
|
|
128
|
+
- Provide next action choices
|
|
129
|
+
- Allow going back or canceling
|
|
130
|
+
|
|
131
|
+
**Context Preservation:**
|
|
132
|
+
- Use diff blocks to show progress
|
|
133
|
+
- Preserve user decisions
|
|
134
|
+
- Allow going back to previous step
|
|
135
|
+
- Enable canceling operations
|
|
136
|
+
|
|
137
|
+
---
|
|
138
|
+
|
|
139
|
+
**Mode management active. Present choices to user.**
|
|
@@ -1,84 +1,84 @@
|
|
|
1
|
-
# Mode Switching Protocol
|
|
2
|
-
|
|
3
|
-
This file contains the detailed instructions for switching Kiro modes. It is referenced by the `/modes {mode_name}` alias in `modes-system.md`.
|
|
4
|
-
|
|
5
|
-
## Mode Switch Steps
|
|
6
|
-
|
|
7
|
-
When switching to mode `{mode_name}`:
|
|
8
|
-
|
|
9
|
-
### Step 1: Load Mode Definition
|
|
10
|
-
|
|
11
|
-
Read `kiro-{mode_name}-mode.md` from agent-system directory into context.
|
|
12
|
-
|
|
13
|
-
This file contains:
|
|
14
|
-
- Mode identity and purpose
|
|
15
|
-
- Interaction protocols
|
|
16
|
-
- Workflow patterns (if any)
|
|
17
|
-
- Response style guidelines
|
|
18
|
-
- Testing approaches
|
|
19
|
-
- File organization rules
|
|
20
|
-
- Integration points
|
|
21
|
-
|
|
22
|
-
### Step 2: Assume Mode Role
|
|
23
|
-
|
|
24
|
-
For this session, you are in **{mode_name} mode**.
|
|
25
|
-
|
|
26
|
-
You will:
|
|
27
|
-
- Follow ALL protocols and instructions from `kiro-{mode_name}-mode.md`
|
|
28
|
-
- Apply mode-specific interaction patterns
|
|
29
|
-
- Use capabilities defined in the mode definition
|
|
30
|
-
- Maintain this mode until user switches modes or ends session
|
|
31
|
-
- Override any conflicting instructions with mode protocols
|
|
32
|
-
|
|
33
|
-
### Step 3: Preserve Context and Check State
|
|
34
|
-
|
|
35
|
-
**IMPORTANT:** Mode switching preserves:
|
|
36
|
-
- All file changes made in previous mode
|
|
37
|
-
- Conversation history and context
|
|
38
|
-
- Current working state
|
|
39
|
-
- User preferences and settings
|
|
40
|
-
|
|
41
|
-
**WARNING:** Mode switching does NOT preserve:
|
|
42
|
-
- Workflow state (e.g., which phase you're in)
|
|
43
|
-
- Approval gate status
|
|
44
|
-
- Phase-specific context
|
|
45
|
-
|
|
46
|
-
**If switching FROM spec mode:**
|
|
47
|
-
Before switching, display warning:
|
|
48
|
-
```
|
|
49
|
-
⚠️ Switching from Spec mode will reset workflow state.
|
|
50
|
-
|
|
51
|
-
Current Spec state:
|
|
52
|
-
- Phase: [current phase]
|
|
53
|
-
- Requirements: [approved/not approved]
|
|
54
|
-
- Design: [approved/not approved]
|
|
55
|
-
- Tasks: [approved/not approved]
|
|
56
|
-
|
|
57
|
-
If you return to Spec mode later, you'll need to re-approve documents.
|
|
58
|
-
|
|
59
|
-
Continue switching to {mode_name} mode? [Yes] [No, stay in Spec mode]
|
|
60
|
-
```
|
|
61
|
-
|
|
62
|
-
Wait for user confirmation before proceeding.
|
|
63
|
-
|
|
64
|
-
You are simply changing HOW you interact, not WHAT you're working on.
|
|
65
|
-
|
|
66
|
-
### Step 4: Begin Interaction
|
|
67
|
-
|
|
68
|
-
Start interaction according to **{mode_name} mode**'s protocols.
|
|
69
|
-
|
|
70
|
-
**If switching to spec mode:**
|
|
71
|
-
- Ask user what they want to work on
|
|
72
|
-
- Offer to create new spec or update existing
|
|
73
|
-
- Follow structured workflow (requirements → design → tasks)
|
|
74
|
-
- Use approval gates between phases
|
|
75
|
-
|
|
76
|
-
**If switching to vibe mode:**
|
|
77
|
-
- Continue with flexible, conversational interaction
|
|
78
|
-
- No formal workflow required
|
|
79
|
-
- Make changes directly when clear
|
|
80
|
-
- Iterate quickly on feedback
|
|
81
|
-
|
|
82
|
-
---
|
|
83
|
-
|
|
84
|
-
**You are now in {mode_name} mode. Begin interaction.**
|
|
1
|
+
# Mode Switching Protocol
|
|
2
|
+
|
|
3
|
+
This file contains the detailed instructions for switching Kiro modes. It is referenced by the `/modes {mode_name}` alias in `modes-system.md`.
|
|
4
|
+
|
|
5
|
+
## Mode Switch Steps
|
|
6
|
+
|
|
7
|
+
When switching to mode `{mode_name}`:
|
|
8
|
+
|
|
9
|
+
### Step 1: Load Mode Definition
|
|
10
|
+
|
|
11
|
+
Read `kiro-{mode_name}-mode.md` from agent-system directory into context.
|
|
12
|
+
|
|
13
|
+
This file contains:
|
|
14
|
+
- Mode identity and purpose
|
|
15
|
+
- Interaction protocols
|
|
16
|
+
- Workflow patterns (if any)
|
|
17
|
+
- Response style guidelines
|
|
18
|
+
- Testing approaches
|
|
19
|
+
- File organization rules
|
|
20
|
+
- Integration points
|
|
21
|
+
|
|
22
|
+
### Step 2: Assume Mode Role
|
|
23
|
+
|
|
24
|
+
For this session, you are in **{mode_name} mode**.
|
|
25
|
+
|
|
26
|
+
You will:
|
|
27
|
+
- Follow ALL protocols and instructions from `kiro-{mode_name}-mode.md`
|
|
28
|
+
- Apply mode-specific interaction patterns
|
|
29
|
+
- Use capabilities defined in the mode definition
|
|
30
|
+
- Maintain this mode until user switches modes or ends session
|
|
31
|
+
- Override any conflicting instructions with mode protocols
|
|
32
|
+
|
|
33
|
+
### Step 3: Preserve Context and Check State
|
|
34
|
+
|
|
35
|
+
**IMPORTANT:** Mode switching preserves:
|
|
36
|
+
- All file changes made in previous mode
|
|
37
|
+
- Conversation history and context
|
|
38
|
+
- Current working state
|
|
39
|
+
- User preferences and settings
|
|
40
|
+
|
|
41
|
+
**WARNING:** Mode switching does NOT preserve:
|
|
42
|
+
- Workflow state (e.g., which phase you're in)
|
|
43
|
+
- Approval gate status
|
|
44
|
+
- Phase-specific context
|
|
45
|
+
|
|
46
|
+
**If switching FROM spec mode:**
|
|
47
|
+
Before switching, display warning:
|
|
48
|
+
```
|
|
49
|
+
⚠️ Switching from Spec mode will reset workflow state.
|
|
50
|
+
|
|
51
|
+
Current Spec state:
|
|
52
|
+
- Phase: [current phase]
|
|
53
|
+
- Requirements: [approved/not approved]
|
|
54
|
+
- Design: [approved/not approved]
|
|
55
|
+
- Tasks: [approved/not approved]
|
|
56
|
+
|
|
57
|
+
If you return to Spec mode later, you'll need to re-approve documents.
|
|
58
|
+
|
|
59
|
+
Continue switching to {mode_name} mode? [Yes] [No, stay in Spec mode]
|
|
60
|
+
```
|
|
61
|
+
|
|
62
|
+
Wait for user confirmation before proceeding.
|
|
63
|
+
|
|
64
|
+
You are simply changing HOW you interact, not WHAT you're working on.
|
|
65
|
+
|
|
66
|
+
### Step 4: Begin Interaction
|
|
67
|
+
|
|
68
|
+
Start interaction according to **{mode_name} mode**'s protocols.
|
|
69
|
+
|
|
70
|
+
**If switching to spec mode:**
|
|
71
|
+
- Ask user what they want to work on
|
|
72
|
+
- Offer to create new spec or update existing
|
|
73
|
+
- Follow structured workflow (requirements → design → tasks)
|
|
74
|
+
- Use approval gates between phases
|
|
75
|
+
|
|
76
|
+
**If switching to vibe mode:**
|
|
77
|
+
- Continue with flexible, conversational interaction
|
|
78
|
+
- No formal workflow required
|
|
79
|
+
- Make changes directly when clear
|
|
80
|
+
- Iterate quickly on feedback
|
|
81
|
+
|
|
82
|
+
---
|
|
83
|
+
|
|
84
|
+
**You are now in {mode_name} mode. Begin interaction.**
|
package/package.json
CHANGED
|
@@ -1,7 +1,15 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "kiro-agents",
|
|
3
|
-
"version": "1.
|
|
4
|
-
"description": "AI agent system for Kiro IDE",
|
|
3
|
+
"version": "1.5.0",
|
|
4
|
+
"description": "Advanced AI agent system for Kiro IDE - Create specialized AI agents, switch interaction modes, and enhance development workflows with minimal cognitive overhead",
|
|
5
|
+
"homepage": "https://github.com/Theadd/kiro-agents#readme",
|
|
6
|
+
"repository": {
|
|
7
|
+
"type": "git",
|
|
8
|
+
"url": "https://github.com/Theadd/kiro-agents.git"
|
|
9
|
+
},
|
|
10
|
+
"bugs": {
|
|
11
|
+
"url": "https://github.com/Theadd/kiro-agents/issues"
|
|
12
|
+
},
|
|
5
13
|
"type": "module",
|
|
6
14
|
"bin": {
|
|
7
15
|
"kiro-agents": "build/npm/bin/cli.js"
|
|
@@ -24,7 +32,22 @@
|
|
|
24
32
|
},
|
|
25
33
|
"keywords": [
|
|
26
34
|
"kiro",
|
|
27
|
-
"
|
|
35
|
+
"kiro-ide",
|
|
36
|
+
"ai-agents",
|
|
37
|
+
"ai-assistant",
|
|
38
|
+
"agent-system",
|
|
39
|
+
"development-tools",
|
|
40
|
+
"ide-extension",
|
|
41
|
+
"ai-workflow",
|
|
42
|
+
"code-assistant",
|
|
43
|
+
"developer-productivity",
|
|
44
|
+
"steering-documents",
|
|
45
|
+
"ai-modes",
|
|
46
|
+
"strict-mode",
|
|
47
|
+
"vibe-mode",
|
|
48
|
+
"spec-mode",
|
|
49
|
+
"typescript",
|
|
50
|
+
"bun"
|
|
28
51
|
],
|
|
29
52
|
"author": "R. Beltran",
|
|
30
53
|
"license": "MIT",
|