@ryodeushii/ai-product-team-agents 0.0.1 → 0.0.3

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 CHANGED
@@ -32,13 +32,13 @@ A collection of agent role definitions, workflow orchestration, and dynamic mode
32
32
  Run from your target project directory:
33
33
 
34
34
  ```bash
35
- bunx github:ryodeushii/ai-product-team-agents <project-name>
35
+ bunx @ryodeushii/ai-product-team-agents <project-name>
36
36
  ```
37
37
 
38
- Or with explicit bin name:
38
+ Or with npm/npx:
39
39
 
40
40
  ```bash
41
- bunx github:ryodeushii/ai-product-team-agents/agents-setup <project-name>
41
+ npx @ryodeushii/ai-product-team-agents <project-name>
42
42
  ```
43
43
 
44
44
  This creates:
@@ -54,11 +54,36 @@ Edit `.agents.yml`:
54
54
 
55
55
  ```yaml
56
56
  project: my-saas
57
- stack: [typescript, react, postgres]
58
- tasks: github-issues # github-issues | linear | files | none
57
+
58
+ # Tech stack used by agents for context
59
+ stack: [typescript, react, postgres, docker]
60
+
61
+ # Task tracking integration
62
+ tasks: github-issues # github-issues | linear | files | none
63
+
64
+ # Docs directory (agents look here for specs, ADRs, etc.)
59
65
  docs_dir: docs/
60
- autonomy: checkpoint # supervised | checkpoint | autonomous
61
- deploy: vercel
66
+
67
+ # How much the orchestrator pauses for human approval
68
+ autonomy: checkpoint # supervised | checkpoint | autonomous
69
+
70
+ # Deployment target (used by devops agent)
71
+ deploy: vercel # vercel | fly | aws | gcp | docker | k8s
72
+
73
+ # Model overrides — all optional, framework defaults apply otherwise
74
+ models:
75
+ defaults:
76
+ thinking: anthropic/claude-sonnet-4-6 # architect, developer, reviewer…
77
+ executing: anthropic/claude-haiku-4-5-20251001 # explorer, fixer, seo
78
+ creative: google/gemini-2.5-pro # designer, marketing
79
+ roles:
80
+ architect: anthropic/claude-sonnet-4-6 # override a specific role
81
+ explorer: local/qwen2.5-coder # local model via OpenCode
82
+
83
+ # Disable roles you don't need
84
+ roles:
85
+ seo: disabled
86
+ marketing: disabled
62
87
  ```
63
88
 
64
89
  ### 3. Use the team
@@ -97,15 +122,17 @@ Models are defined in `models.yml`. Three tiers:
97
122
 
98
123
  ### Per-project overrides
99
124
 
100
- In your project's `.agents.yml`:
125
+ In your project's `.agents.yml`, override any tier or individual role:
101
126
 
102
127
  ```yaml
103
128
  models:
104
129
  defaults:
105
- creative: google/gemini-2.5-pro # override a whole tier
130
+ thinking: anthropic/claude-sonnet-4-6 # all thinking-tier roles
131
+ executing: anthropic/claude-haiku-4-5-20251001 # all executing-tier roles
132
+ creative: google/gemini-2.5-pro # all creative-tier roles
106
133
  roles:
107
- architect: openai/gpt-4.1 # override a specific role
108
- explorer: local/qwen2.5-coder # local model via OpenCode
134
+ architect: openai/gpt-4.1 # one specific role
135
+ explorer: local/qwen2.5-coder # local model via OpenCode
109
136
  ```
110
137
 
111
138
  Resolution order (highest to lowest priority):
@@ -118,13 +145,33 @@ Models are resolved **dynamically** — changes to `models.yml` or `.agents.yml`
118
145
 
119
146
  ### MCP server
120
147
 
121
- The framework ships an MCP server that exposes a `resolve_model(role, project_root)` tool for dynamic model resolution. Run it with:
148
+ The framework ships an MCP server that exposes a `resolve_model(role, project_root)` tool for dynamic model resolution.
122
149
 
123
- ```bash
124
- bun run src/mcp/server.ts
150
+ Add to your Claude Code MCP config (`~/.claude/mcp.json` or project-level `.mcp.json`):
151
+
152
+ ```json
153
+ {
154
+ "mcpServers": {
155
+ "agents-framework": {
156
+ "command": "bunx",
157
+ "args": ["@ryodeushii/ai-product-team-agents/agents-mcp"]
158
+ }
159
+ }
160
+ }
125
161
  ```
126
162
 
127
- Add to your Claude Code or OpenCode MCP config to enable per-project model overrides.
163
+ Or with npx:
164
+
165
+ ```json
166
+ {
167
+ "mcpServers": {
168
+ "agents-framework": {
169
+ "command": "npx",
170
+ "args": ["-y", "@ryodeushii/ai-product-team-agents/agents-mcp"]
171
+ }
172
+ }
173
+ }
174
+ ```
128
175
 
129
176
  ---
130
177
 
@@ -162,8 +209,8 @@ New projects always pause at every step regardless of this setting.
162
209
  ### Setup
163
210
 
164
211
  ```bash
165
- git clone <repo>
166
- cd agents
212
+ git clone https://github.com/ryodeushii/ai-product-team-agents
213
+ cd ai-product-team-agents
167
214
  bun install
168
215
  ```
169
216
 
package/package.json CHANGED
@@ -1,6 +1,10 @@
1
1
  {
2
+ "$schema": "https://json.schemastore.org/package.json",
2
3
  "name": "@ryodeushii/ai-product-team-agents",
3
- "version": "0.0.1",
4
+ "version": "0.0.3",
5
+ "repository": {
6
+ "url": "https://github.com/ryodeushii/ai-product-team-agents"
7
+ },
4
8
  "module": "index.ts",
5
9
  "type": "module",
6
10
  "files": [
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  name: orchestrator
3
3
  description: Delegates to product team specialists for optimal quality, speed, and cost
4
- mode: agent
4
+ mode: primary
5
5
  ---
6
6
 
7
7
  <Role>
@@ -1,17 +1,56 @@
1
1
  project: {{project_name}}
2
+
3
+ # Tech stack — used by agents for context
2
4
  stack: []
5
+ # stack: [typescript, react, postgres, docker]
6
+
7
+ # Task tracking integration
3
8
  tasks: none # github-issues | linear | files | none
9
+
10
+ # Docs directory (agents look here for specs, ADRs, etc.)
4
11
  docs_dir: docs/
12
+
13
+ # How much the orchestrator pauses for human approval
5
14
  autonomy: checkpoint # supervised | checkpoint | autonomous
6
- # deploy: vercel
15
+ # supervised — pause before every agent handoff
16
+ # checkpoint — pause at key decisions (spec, arch, deploy)
17
+ # autonomous — pause only on errors or blockers
18
+
19
+ # Deployment target (used by devops agent)
20
+ # deploy: vercel # vercel | fly | aws | gcp | docker | k8s
7
21
 
22
+ # ─── Model overrides ────────────────────────────────────────────────────────
23
+ #
24
+ # Resolution order (highest → lowest priority):
25
+ # 1. models.roles.<role> — this project, specific role
26
+ # 2. models.defaults.<tier> — this project, whole tier
27
+ # 3. framework roles default — models.yml roles section
28
+ # 4. framework tier default — models.yml defaults section
29
+ #
8
30
  # models:
9
31
  # defaults:
10
- # creative: google/gemini-2.5-pro
32
+ # thinking: anthropic/claude-sonnet-4-6 # architect, developer, reviewer…
33
+ # executing: anthropic/claude-haiku-4-5-20251001 # explorer, fixer, seo
34
+ # creative: google/gemini-2.5-pro # designer, marketing
11
35
  # roles:
12
- # architect: openai/gpt-4.1
13
- # explorer: local/qwen2.5-coder
36
+ # orchestrator: anthropic/claude-sonnet-4-6
37
+ # architect: anthropic/claude-sonnet-4-6
38
+ # developer: anthropic/claude-sonnet-4-6
39
+ # reviewer: anthropic/claude-sonnet-4-6
40
+ # qa: anthropic/claude-sonnet-4-6
41
+ # devops: anthropic/claude-sonnet-4-6
42
+ # pm: anthropic/claude-sonnet-4-6
43
+ # explorer: anthropic/claude-haiku-4-5-20251001
44
+ # fixer: anthropic/claude-haiku-4-5-20251001
45
+ # seo: anthropic/claude-haiku-4-5-20251001
46
+ # designer: google/gemini-2.5-pro
47
+ # marketing: google/gemini-2.5-pro
48
+ # # local models via OpenCode:
49
+ # # explorer: local/qwen2.5-coder
50
+ # # fixer: local/qwen2.5-coder
14
51
 
52
+ # ─── Disable roles you don't need ───────────────────────────────────────────
15
53
  # roles:
16
- # seo: disabled
54
+ # seo: disabled
17
55
  # marketing: disabled
56
+ # designer: disabled