design-review-gate

Name: design-review-gate
Author: dsifry

361stars51forks

How to install design-review-gate?

dsifry/metaswarm/design-review-gate

$npx -y skills add dsifry/metaswarm --skill design-review-gate

Installs into the current project.

Summary

Automatic review gate that runs after brainstorming completes - spawns PM, Architect, Designer, Security, and CTO agents in parallel, iterates until all approve

Files · 1

View on GitHub

SKILL.md

1	# Design Review Gate
2
3	## Purpose
4
5	This skill automatically activates after a design document is created (typically via `superpowers:brainstorming`). It ensures complex features receive proper review before implementation begins by spawning five specialist agents:
6
7	1. Product Manager Agent - Use case validation and user benefit review
8	2. Architect Agent - Technical architecture review
9	3. Designer Agent - UX/API design quality review
10	4. Security Design Agent - Security threat modeling and protection review
11	5. CTO Agent - Codebase alignment and TDD readiness review
12
13	All five must approve before proceeding to implementation.
14
15	---
16
17	## Coordination Mode Note
18
19	This skill supports both coordination modes:
20
21	- Task Mode (default): Spawn 5 parallel `Task()` subagents for each review round. Fresh instances per round.
22	- Team Mode: `TeamCreate("review-{design-doc-name}")`, spawn 5 reviewers as named teammates (`pm`, `architect`, `designer`, `security`, `cto`). Reviewers retain context through revision cycles — saves 5 cold starts per re-review round. After approval, send `shutdown_request` to all, then `TeamDelete`.
23
24	In either mode, the review criteria, iteration protocol, and escalation rules are identical. See `./guides/agent-coordination.md` for mode detection.
25
26	---
27
28	## Activation Triggers
29
30	This skill auto-activates when:
31
32	1. A design document is committed to `docs/plans/*-design.md`
33	2. The `superpowers:brainstorming` skill completes
34	3. User explicitly requests: `/review-design <path-to-design.md>`
35
36	---
37
38	## Workflow
39
40	### Phase 1: Spawn Review Agents (Parallel)
41
42	```typescript
43	// Spawn all five agents in parallel for efficiency
44	const [pmResult, architectResult, designerResult, securityResult, ctoResult] = await Promise.all([
45	Task({
46	subagent_type: "general-purpose",
47	description: "PM review",
48	prompt: pmReviewPrompt(designDocPath),
49	}),
50	Task({
51	subagent_type: "general-purpose",
52	description: "Architect review",
53	prompt: architectReviewPrompt(designDocPath),
54	}),
55	Task({
56	subagent_type: "general-purpose",
57	description: "Designer review",
58	prompt: designerReviewPrompt(designDocPath),
59	}),
60	Task({
61	subagent_type: "general-purpose",
62	description: "Security design review",
63	prompt: securityDesignReviewPrompt(designDocPath),
64	}),
65	Task({
66	subagent_type: "general-purpose",
67	description: "CTO review",
68	prompt: ctoReviewPrompt(designDocPath),
69	}),
70	]);
71	```
72
73	### Phase 2: Aggregate Results
74
75	Each agent returns a structured review:
76
77	```typescript
78	interface ReviewResult {
79	agent: "product-manager" \| "architect" \| "designer" \| "security-design" \| "cto";
80	verdict: "APPROVED" \| "NEEDS_REVISION";
81	blockers: string[]; // MUST fix before implementation
82	suggestions: string[]; // Nice to have
83	questions: string[]; // Clarifications needed
84	use_case_analysis?: {
85	// PM agent only
86	total_use_cases: number;
87	clear: number;
88	needs_work: number;
89	missing_scenarios: string[];
90	};
91	threat_model?: {
92	// Security agent only
93	high_risk: string[];
94	medium_risk: string[];
95	mitigations_required: string[];
96	};
97	}
98	```
99
100	### Phase 3: Check Gate
101
102	```
103	IF all five agents return APPROVED:
104	→ Proceed to implementation
105	→ Create epic with approved design
106
107	ELSE:
108	→ Consolidate feedback
109	→ Present to user
110	→ Iterate on design
111	→ Re-run gate (max 3 iterations)
112	```
113
114	### Phase 4: Human Escalation
115
116	After 3 failed iterations:
117
118	1. Create summary of remaining blockers
119	2. Ask human to decide: override, defer, or cancel
120
121	---
122
123	## Agent Prompts
124
125	### Product Manager Agent Prompt
126
127	````markdown
128	You are the PRODUCT MANAGER AGENT reviewing a design document.
129
130	## Your Task
131
132	Review this design for use case clarity and user benefit validation.
133
134	## Design Document
135
136	<path>: {designDocPath}
137
138	## Review Criteria
139
140	### 1. Use Case Validation
141
142	- [ ] Each use case follows WHO/WANTS/SO THAT/WHEN format
143	- [ ] User personas are clearly defined
144	- [ ] Use cases are realistic and based on user needs
145	- [ ] Edge cases and error scenarios considered from user perspective
146
147	### 2. User Benefit Review
148
149	- [ ] Value proposition clearly articulated
150	- [ ] Benefits are measurable (time saved, success rate, etc.)
151	- [ ] User journey impact understood
152	- [ ] Improvement is significant enough to build
153
154	### 3. Scope Assessment
155
156	- [ ] MVP clearly defined (must have vs nice to have)
157	- [ ] No feature creep detected
158	- [ ] Scope matches user needs, not technical possibilities
159	- [ ] "Solution looking for problem" anti-pattern avoided
160
161	### 4. Success Criteria
162
163	- [ ] Success metrics are user-focused (not just technical)
164	- [ ] Metrics are measurable and have thresh

People who liked this also viewed

6 picks

Type

requesting-code-reviewUse when completing tasks, implementing major features, or before merging to verify work meets requirementsSkillsJul 2026176k259k
code-reviewSkills for Real Engineers. Straight from my .agents directory.SkillsJul 2026158k181k
receiving-code-reviewUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performativeSkillsJul 2026146k259k
baoyu-format-markdownFormats plain text or markdown files with frontmatter, titles, summaries, headings, bold, lists, and code blocks. Use when user asks to "format markdown", "beautify article", "add formatting", or imprSkillsJul 202625k24k
code-review-excellenceMaster effective code review practices to provide constructive feedback, catch bugs early, and foster knowledge sharing while maintaining team morale. Use when reviewing pull requests, establishing reSkillsJul 202625k38k
code-review-and-qualityConducts multi-axis code review. Use before merging any change. Use when reviewing code written by yourself, another agent, or a human. Use when you need to assess code quality across multiple dimensiSkillsJul 202618k80k

design-review-gate

361stars51forks

View on GitHub

How to install design-review-gate?

dsifry/metaswarm/design-review-gate

$npx -y skills add dsifry/metaswarm --skill design-review-gate

Installs into the current project.

Summary

Automatic review gate that runs after brainstorming completes - spawns PM, Architect, Designer, Security, and CTO agents in parallel, iterates until all approve

Files · 1

View on GitHub

SKILL.md

1	# Design Review Gate
2
3	## Purpose
4
5	This skill automatically activates after a design document is created (typically via `superpowers:brainstorming`). It ensures complex features receive proper review before implementation begins by spawning five specialist agents:
6
7	1. Product Manager Agent - Use case validation and user benefit review
8	2. Architect Agent - Technical architecture review
9	3. Designer Agent - UX/API design quality review
10	4. Security Design Agent - Security threat modeling and protection review
11	5. CTO Agent - Codebase alignment and TDD readiness review
12
13	All five must approve before proceeding to implementation.
14
15	---
16
17	## Coordination Mode Note
18
19	This skill supports both coordination modes:
20
21	- Task Mode (default): Spawn 5 parallel `Task()` subagents for each review round. Fresh instances per round.
22	- Team Mode: `TeamCreate("review-{design-doc-name}")`, spawn 5 reviewers as named teammates (`pm`, `architect`, `designer`, `security`, `cto`). Reviewers retain context through revision cycles — saves 5 cold starts per re-review round. After approval, send `shutdown_request` to all, then `TeamDelete`.
23
24	In either mode, the review criteria, iteration protocol, and escalation rules are identical. See `./guides/agent-coordination.md` for mode detection.
25
26	---
27
28	## Activation Triggers
29
30	This skill auto-activates when:
31
32	1. A design document is committed to `docs/plans/*-design.md`
33	2. The `superpowers:brainstorming` skill completes
34	3. User explicitly requests: `/review-design <path-to-design.md>`
35
36	---
37
38	## Workflow
39
40	### Phase 1: Spawn Review Agents (Parallel)
41
42	```typescript
43	// Spawn all five agents in parallel for efficiency
44	const [pmResult, architectResult, designerResult, securityResult, ctoResult] = await Promise.all([
45	Task({
46	subagent_type: "general-purpose",
47	description: "PM review",
48	prompt: pmReviewPrompt(designDocPath),
49	}),
50	Task({
51	subagent_type: "general-purpose",
52	description: "Architect review",
53	prompt: architectReviewPrompt(designDocPath),
54	}),
55	Task({
56	subagent_type: "general-purpose",
57	description: "Designer review",
58	prompt: designerReviewPrompt(designDocPath),
59	}),
60	Task({
61	subagent_type: "general-purpose",
62	description: "Security design review",
63	prompt: securityDesignReviewPrompt(designDocPath),
64	}),
65	Task({
66	subagent_type: "general-purpose",
67	description: "CTO review",
68	prompt: ctoReviewPrompt(designDocPath),
69	}),
70	]);
71	```
72
73	### Phase 2: Aggregate Results
74
75	Each agent returns a structured review:
76
77	```typescript
78	interface ReviewResult {
79	agent: "product-manager" \| "architect" \| "designer" \| "security-design" \| "cto";
80	verdict: "APPROVED" \| "NEEDS_REVISION";
81	blockers: string[]; // MUST fix before implementation
82	suggestions: string[]; // Nice to have
83	questions: string[]; // Clarifications needed
84	use_case_analysis?: {
85	// PM agent only
86	total_use_cases: number;
87	clear: number;
88	needs_work: number;
89	missing_scenarios: string[];
90	};
91	threat_model?: {
92	// Security agent only
93	high_risk: string[];
94	medium_risk: string[];
95	mitigations_required: string[];
96	};
97	}
98	```
99
100	### Phase 3: Check Gate
101
102	```
103	IF all five agents return APPROVED:
104	→ Proceed to implementation
105	→ Create epic with approved design
106
107	ELSE:
108	→ Consolidate feedback
109	→ Present to user
110	→ Iterate on design
111	→ Re-run gate (max 3 iterations)
112	```
113
114	### Phase 4: Human Escalation
115
116	After 3 failed iterations:
117
118	1. Create summary of remaining blockers
119	2. Ask human to decide: override, defer, or cancel
120
121	---
122
123	## Agent Prompts
124
125	### Product Manager Agent Prompt
126
127	````markdown
128	You are the PRODUCT MANAGER AGENT reviewing a design document.
129
130	## Your Task
131
132	Review this design for use case clarity and user benefit validation.
133
134	## Design Document
135
136	<path>: {designDocPath}
137
138	## Review Criteria
139
140	### 1. Use Case Validation
141
142	- [ ] Each use case follows WHO/WANTS/SO THAT/WHEN format
143	- [ ] User personas are clearly defined
144	- [ ] Use cases are realistic and based on user needs
145	- [ ] Edge cases and error scenarios considered from user perspective
146
147	### 2. User Benefit Review
148
149	- [ ] Value proposition clearly articulated
150	- [ ] Benefits are measurable (time saved, success rate, etc.)
151	- [ ] User journey impact understood
152	- [ ] Improvement is significant enough to build
153
154	### 3. Scope Assessment
155
156	- [ ] MVP clearly defined (must have vs nice to have)
157	- [ ] No feature creep detected
158	- [ ] Scope matches user needs, not technical possibilities
159	- [ ] "Solution looking for problem" anti-pattern avoided
160
161	### 4. Success Criteria
162
163	- [ ] Success metrics are user-focused (not just technical)
164	- [ ] Metrics are measurable and have thresh

People who liked this also viewed

6 picks

Type