codingbuddy-rules 0.0.0-canary.20260106061736.7e8c714 → 0.0.0-canary.20260106081656.f966936
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.
|
@@ -165,3 +165,95 @@ recommend_skills({ prompt: "Build a dashboard component" })
|
|
|
165
165
|
**Supported Languages:** English, Korean, Japanese, Chinese, Spanish
|
|
166
166
|
|
|
167
167
|
The tool returns skill recommendations with confidence levels (high/medium) and matched patterns for transparency.
|
|
168
|
+
|
|
169
|
+
## Agent Hierarchy
|
|
170
|
+
|
|
171
|
+
CodingBuddy uses a layered agent hierarchy for different types of tasks:
|
|
172
|
+
|
|
173
|
+
### Tier 1: Primary Agents (Mode-specific)
|
|
174
|
+
|
|
175
|
+
| Mode | Agents | Description |
|
|
176
|
+
|------|--------|-------------|
|
|
177
|
+
| **PLAN** | solution-architect, technical-planner | Design and planning tasks |
|
|
178
|
+
| **ACT** | frontend-developer, backend-developer, devops-engineer, agent-architect | Implementation tasks |
|
|
179
|
+
| **EVAL** | code-reviewer | Code review and evaluation |
|
|
180
|
+
|
|
181
|
+
### Tier 2: Specialist Agents
|
|
182
|
+
|
|
183
|
+
Specialist agents can be invoked by any Primary Agent as needed:
|
|
184
|
+
|
|
185
|
+
- security-specialist
|
|
186
|
+
- accessibility-specialist
|
|
187
|
+
- performance-specialist
|
|
188
|
+
- test-strategy-specialist
|
|
189
|
+
- documentation-specialist
|
|
190
|
+
- architecture-specialist
|
|
191
|
+
- code-quality-specialist
|
|
192
|
+
- seo-specialist
|
|
193
|
+
- design-system-specialist
|
|
194
|
+
|
|
195
|
+
### Agent Resolution
|
|
196
|
+
|
|
197
|
+
1. **PLAN mode**: Always uses `solution-architect` or `technical-planner` based on prompt analysis
|
|
198
|
+
2. **ACT mode**: Uses recommended agent from PLAN, or falls back to AI analysis
|
|
199
|
+
3. **EVAL mode**: Always uses `code-reviewer`
|
|
200
|
+
|
|
201
|
+
## Activation Messages
|
|
202
|
+
|
|
203
|
+
When agents or skills are activated, CodingBuddy displays activation messages for transparency:
|
|
204
|
+
|
|
205
|
+
### Output Format
|
|
206
|
+
|
|
207
|
+
```
|
|
208
|
+
🤖 solution-architect [Primary Agent]
|
|
209
|
+
👤 security-specialist [Specialist] (by solution-architect)
|
|
210
|
+
⚡ brainstorming [Specialist] (by technical-planner)
|
|
211
|
+
```
|
|
212
|
+
|
|
213
|
+
### Icons
|
|
214
|
+
|
|
215
|
+
| Icon | Meaning |
|
|
216
|
+
|------|---------|
|
|
217
|
+
| 🤖 | Primary Agent |
|
|
218
|
+
| 👤 | Specialist Agent |
|
|
219
|
+
| ⚡ | Skill |
|
|
220
|
+
|
|
221
|
+
### ParseMode Response Fields
|
|
222
|
+
|
|
223
|
+
The `parse_mode` MCP tool returns these agent-related fields:
|
|
224
|
+
|
|
225
|
+
```json
|
|
226
|
+
{
|
|
227
|
+
"mode": "PLAN",
|
|
228
|
+
"delegates_to": "solution-architect",
|
|
229
|
+
"primary_agent_source": "intent",
|
|
230
|
+
"activation_message": {
|
|
231
|
+
"formatted": "🤖 solution-architect [Primary Agent]",
|
|
232
|
+
"activations": [
|
|
233
|
+
{
|
|
234
|
+
"type": "agent",
|
|
235
|
+
"name": "solution-architect",
|
|
236
|
+
"tier": "primary",
|
|
237
|
+
"timestamp": "2024-01-06T12:00:00Z"
|
|
238
|
+
}
|
|
239
|
+
]
|
|
240
|
+
},
|
|
241
|
+
"recommended_act_agent": {
|
|
242
|
+
"agentName": "backend-developer",
|
|
243
|
+
"reason": "API implementation task detected",
|
|
244
|
+
"confidence": 0.9
|
|
245
|
+
}
|
|
246
|
+
}
|
|
247
|
+
```
|
|
248
|
+
|
|
249
|
+
### Displaying Activation Messages
|
|
250
|
+
|
|
251
|
+
AI assistants should display the `activation_message.formatted` field at the start of their response:
|
|
252
|
+
|
|
253
|
+
```
|
|
254
|
+
🤖 solution-architect [Primary Agent]
|
|
255
|
+
|
|
256
|
+
# Mode: PLAN
|
|
257
|
+
|
|
258
|
+
...
|
|
259
|
+
```
|
package/package.json
CHANGED