greenfield

Name: greenfield
Author: wednesday-solutions

161stars19forks

How to install greenfield?

wednesday-solutions/ai-agent-skills/greenfield

$npx -y skills add wednesday-solutions/ai-agent-skills --skill greenfield

Installs into the current project.

Summary

Parallel persona planning for new projects. Research agent runs first to build domain context, then Architect, PM, and Security agents run in parallel. Synthesis agent combines all perspectives into a detailed GSD-style PLAN.md with Tensions section.

Files · 1

View on GitHub

SKILL.md

1	# Greenfield Planning Skill
2
3	## Trigger
4
5	Run once per project: `ws-skills plan`
6
7	Reads `BRIEF.md` from the project root (or prompts for one). Asks 5 clarifying questions before planning.
8
9	## Flow
10
11	```
12	Brief + Q&A
13	↓
14	Research agent (sequential) ← domain landscape, ecosystem, hidden complexity
15	↓
16	┌─────────────────────────────────────┐
17	│ Architect │ PM │ Security (parallel)│ ← spawn 3 subagents simultaneously
18	└─────────────────────────────────────┘
19	↓
20	Synthesis ← combines all into PLAN.md
21	```
22
23	## Agents
24
25	### 1. Research (sequential — runs first)
26
27	Builds domain context that all other agents receive. Covers:
28	- Existing solutions and their weaknesses
29	- Standard and emerging tech stacks for this domain
30	- Technologies to avoid and why
31	- Non-obvious domain challenges
32	- Integration landscape (auth, payments, comms, etc.)
33	- Regulatory and compliance context
34	- Realistic timeline based on similar projects
35	- Hidden complexity — things that take 3x longer than expected
36	- Success patterns from the best products in this space
37
38	Output: `research.md`
39
40	### 2–4. Architect, PM, Security (parallel subagents)
41
42	Spawn all three simultaneously using the Agent tool. Each receives the full brief, Q&A, and research output as context.
43
44	```
45	Agent 1 — Architect
46	Agent 2 — PM ← launch all three in a single message, do not wait
47	Agent 3 — Security
48	```
49
50	Wait for all three to complete before running Synthesis.
51
52	Architect output: `architect.md`
53	- System design overview
54	- Tech stack with rationale per layer
55	- Module boundaries and interfaces
56	- Infrastructure and CI/CD
57	- Scaling strategy
58	- Technical risks
59
60	PM output: `pm.md`
61	- Phases with tasks and acceptance criteria
62	- Success metrics
63	- Out of scope items
64	- Assumptions
65
66	Security output: `security.md`
67	- Threat model (likelihood + impact)
68	- Data classification
69	- Auth strategy recommendation
70	- Compliance flags
71	- Concrete security tasks
72	- Urgent flags
73
74	### 5. Synthesis
75
76	Combines research + all three persona outputs into a single PLAN.md covering:
77	- Overview
78	- Clarifications table
79	- Tech stack
80	- Architecture
81	- Phases with tasks and acceptance criteria
82	- Security plan
83	- Success metrics
84	- Risks
85	- Tensions (unresolved disagreements between personas)
86	- Assumptions
87	- Out of scope
88	- Branch naming (GIT-OS format)
89
90	Output: `PLAN.md`
91
92	## Tools
93
94	\| Action \| Tool \|
95	\|--------\|------\|
96	\| Read `BRIEF.md` \| `Read` \|
97	\| Write persona output files (`architect.md`, `pm.md`, etc.) \| `Write` \|
98	\| Spawn Architect, PM, Security personas in parallel \| `Agent` (3 calls in one message) \|
99	\| Search the brief for keywords \| `Grep` \|
100
101	## Output Location
102
103	All files written to `.wednesday/plans/` in the target directory:
104
105	```
106	.wednesday/plans/
107	├── research.md ← domain context
108	├── architect.md ← technical design
109	├── pm.md ← phases and metrics
110	├── security.md ← threat model
111	└── PLAN.md ← combined PRD (primary output)
112	```
113
114	## Failure Handling
115
116	Each agent fails independently. If one fails, the others continue and synthesis runs with whatever data is available. Failed agents show `[partial fallback]` in the progress display.
117
118	## Rules
119
120	- Branch naming in PLAN.md must follow GIT-OS format
121	- Never generate `CODEBASE.md` for greenfield projects — it doesn't exist yet
122	- Cost target: under $0.20 per run

People who liked this also viewed

6 picks

Type

greenfield

161stars19forks

View on GitHub

How to install greenfield?

wednesday-solutions/ai-agent-skills/greenfield

$npx -y skills add wednesday-solutions/ai-agent-skills --skill greenfield

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Greenfield Planning Skill
2
3	## Trigger
4
5	Run once per project: `ws-skills plan`
6
7	Reads `BRIEF.md` from the project root (or prompts for one). Asks 5 clarifying questions before planning.
8
9	## Flow
10
11	```
12	Brief + Q&A
13	↓
14	Research agent (sequential) ← domain landscape, ecosystem, hidden complexity
15	↓
16	┌─────────────────────────────────────┐
17	│ Architect │ PM │ Security (parallel)│ ← spawn 3 subagents simultaneously
18	└─────────────────────────────────────┘
19	↓
20	Synthesis ← combines all into PLAN.md
21	```
22
23	## Agents
24
25	### 1. Research (sequential — runs first)
26
27	Builds domain context that all other agents receive. Covers:
28	- Existing solutions and their weaknesses
29	- Standard and emerging tech stacks for this domain
30	- Technologies to avoid and why
31	- Non-obvious domain challenges
32	- Integration landscape (auth, payments, comms, etc.)
33	- Regulatory and compliance context
34	- Realistic timeline based on similar projects
35	- Hidden complexity — things that take 3x longer than expected
36	- Success patterns from the best products in this space
37
38	Output: `research.md`
39
40	### 2–4. Architect, PM, Security (parallel subagents)
41
42	Spawn all three simultaneously using the Agent tool. Each receives the full brief, Q&A, and research output as context.
43
44	```
45	Agent 1 — Architect
46	Agent 2 — PM ← launch all three in a single message, do not wait
47	Agent 3 — Security
48	```
49
50	Wait for all three to complete before running Synthesis.
51
52	Architect output: `architect.md`
53	- System design overview
54	- Tech stack with rationale per layer
55	- Module boundaries and interfaces
56	- Infrastructure and CI/CD
57	- Scaling strategy
58	- Technical risks
59
60	PM output: `pm.md`
61	- Phases with tasks and acceptance criteria
62	- Success metrics
63	- Out of scope items
64	- Assumptions
65
66	Security output: `security.md`
67	- Threat model (likelihood + impact)
68	- Data classification
69	- Auth strategy recommendation
70	- Compliance flags
71	- Concrete security tasks
72	- Urgent flags
73
74	### 5. Synthesis
75
76	Combines research + all three persona outputs into a single PLAN.md covering:
77	- Overview
78	- Clarifications table
79	- Tech stack
80	- Architecture
81	- Phases with tasks and acceptance criteria
82	- Security plan
83	- Success metrics
84	- Risks
85	- Tensions (unresolved disagreements between personas)
86	- Assumptions
87	- Out of scope
88	- Branch naming (GIT-OS format)
89
90	Output: `PLAN.md`
91
92	## Tools
93
94	\| Action \| Tool \|
95	\|--------\|------\|
96	\| Read `BRIEF.md` \| `Read` \|
97	\| Write persona output files (`architect.md`, `pm.md`, etc.) \| `Write` \|
98	\| Spawn Architect, PM, Security personas in parallel \| `Agent` (3 calls in one message) \|
99	\| Search the brief for keywords \| `Grep` \|
100
101	## Output Location
102
103	All files written to `.wednesday/plans/` in the target directory:
104
105	```
106	.wednesday/plans/
107	├── research.md ← domain context
108	├── architect.md ← technical design
109	├── pm.md ← phases and metrics
110	├── security.md ← threat model
111	└── PLAN.md ← combined PRD (primary output)
112	```
113
114	## Failure Handling
115
116	Each agent fails independently. If one fails, the others continue and synthesis runs with whatever data is available. Failed agents show `[partial fallback]` in the progress display.
117
118	## Rules
119
120	- Branch naming in PLAN.md must follow GIT-OS format
121	- Never generate `CODEBASE.md` for greenfield projects — it doesn't exist yet
122	- Cost target: under $0.20 per run

People who liked this also viewed

6 picks

Type