co-design

Name: co-design
Author: am-will

222stars12forks

How to install co-design?

am-will/swarms/co-design

$npx -y skills add am-will/swarms --skill co-design

Installs into the current project.

Summary

Only to be triggered by explicit /co-design commands. Parallel task executor that routes frontend/design tasks to Claude CLI print mode for superior design output.

Files · 1

View on GitHub

SKILL.md

1	# Co-Design: Parallel Task Executor with Design-Mode Routing
2
3	You are an Orchestrator for subagents. Use orchestration mode to parse plan files and delegate tasks to parallel subagents using task dependencies, in a loop, until all tasks are completed.
4
5	Key difference from standard parallel-task: When a task involves frontend design, styling, UI, or visual work, you launch it via `claude -p` (print mode) in a background shell instead of using the Task tool. This gives the design agent full CLI capabilities and tool access. All other tasks use normal Task tool subagents.
6
7	## Task Classification
8
9	Before launching each task, classify it as design or standard:
10
11	### Design Tasks (route to `claude -p`)
12	A task is a design task if it primarily involves ANY of:
13	- CSS, SCSS, Tailwind, or any styling
14	- HTML structure/templates for UI
15	- React/Vue/Svelte/Angular component markup and styling
16	- Layout, responsive design, grid, flexbox
17	- Design tokens, theme files, color schemes
18	- UI component creation (buttons, cards, modals, navbars, etc.)
19	- Animation, transitions, visual effects
20	- Accessibility (a11y) related to visual presentation
21	- Asset management (icons, images, fonts)
22	- Design system implementation
23
24	### Standard Tasks (route to Task tool subagent)
25	Everything else:
26	- Backend logic, API endpoints, database
27	- Business logic, services, utilities
28	- Configuration, tooling, CI/CD
29	- Testing (unless it's visual/snapshot testing)
30	- State management, data fetching
31	- Authentication, authorization
32
33	When in doubt: If a task mixes both (e.g., "create a form with validation"), route it as a design task since it has a UI component.
34
35	## Process
36
37	### Step 1: Parse Request
38
39	Extract from user request:
40	1. Plan file: The markdown plan to read
41	2. Task subset (optional): Specific task IDs to run
42
43	If no subset provided, run the full plan.
44
45	### Step 2: Read & Parse Plan
46
47	1. Find task subsections (e.g., `### T1:` or `### Task 1.1:`)
48	2. For each task, extract:
49	- Task ID and name
50	- depends_on list (from `- depends_on: [...]`)
51	- Full content (description, location, acceptance criteria, validation)
52	3. Build task list
53	4. Classify each task as `design` or `standard` based on the rules above
54	5. If a task subset was requested, filter the task list to only those IDs and their required dependencies.
55
56	### Step 3: Launch Tasks
57
58	For each unblocked task, launch it based on its classification:
59
60	#### For STANDARD tasks → Use Task tool subagent
61
62	Launch with:
63	- description: "Implement task [ID]: [name]"
64	- prompt: Use the Task Prompt Template below
65	- subagent_type: "parallel"
66
67	#### For DESIGN tasks → Use `claude -p` in background shell
68
69	Construct the exact same prompt you would give a subagent, then launch via Bash:
70
71	```bash
72	claude -p "YOUR_PROMPT_HERE" \
73	--allowedTools "Bash,Read,Edit,Write,Glob,Grep,WebFetch,WebSearch" \
74	--max-turns 50 \
75	--output-format text \
76	> /tmp/co-design-[TASK_ID]-output.log 2>&1 &
77	```
78
79	CRITICAL rules for design task launch:
80	1. The prompt content must be IDENTICAL to what you'd give a Task tool subagent (use the same Task Prompt Template)
81	2. Escape the prompt properly for shell execution - use a heredoc approach:
82	```bash
83	claude -p "$(cat <<'PROMPT_EOF'
84	[full prompt content here]
85	PROMPT_EOF
86	)" \
87	--allowedTools "Bash,Read,Edit,Write,Glob,Grep,WebFetch,WebSearch" \
88	--max-turns 50 \
89	--output-format text \
90	> /tmp/co-design-[TASK_ID]-output.log 2>&1 &
91	```
92	3. Log output to `/tmp/co-design-[TASK_ID]-output.log` for later inspection
93	4. Track the background PID so you can check completion
94	5. Launch ALL unblocked design tasks in parallel (multiple background shells)
95
96	Launch all unblocked tasks (both types) in parallel. A task is unblocked if all IDs in its depends_on list are complete.
97
98	### Task Prompt Template
99
100	Use this for BOTH standard subagents AND design-mode `claude -p` tasks:
101
102	```
103	You are implementing a specific task from a development plan.
104
105	## Context
106	- Plan: [filename]
107	- Goals: [relevant overview from plan]
108	- Dependencies: [prerequisites for this task]
109	- Related tasks: [tasks that depend on or are depended on by this task]
110	- Constraints: [risks from plan]
111
112	## Your Task
113	Task [ID]: [Name]
114
115	Location: [File paths]
116	Description: [Full description]
117
118	Acceptance Criteria:
119	[List from plan]
120
121	Validation:
122	[Tests or verification from plan]
123
124	## Instructions
125	1. Read the working plan and fully understand this task before coding.
126	2. Read all relevant files first, then do targeted codebase research (related modules, tests, call sites, and dependencies) to confirm the approach.
127	3. Default to TDD RED phase first using a `tdd_test_writer` subagent:
128	- Pass task context and acceptance criteria.
129	- Require tests-only edits.

People who liked this also viewed

6 picks

Type

paperclip-create-agentCreate new agents in Paperclip with governance-aware hiring. Use when you need to inspect adapter configuration options, compare existing agent configs, draft a new agent prompt/config, and submit a hSkillsJul 2026157k5
subagent-driven-developmentUse when executing implementation plans with independent tasks in the current sessionSkillsJul 2026154k259k
dispatching-parallel-agentsUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesSkillsJul 2026143k259k
browser-act-skill-forgeForges reusable Skill packages (SKILL.md + scripts) from website exploration via browser-act — no re-exploration later. Use when: user wants a reusable Skill for any website, needs to understand a sitSkillsJul 202685k4.6k
wizardGenerate an interactive bash wizard that walks a human through a manual procedure — third-party setup, a one-off migration, an A→B state transition — opening URLs, capturing values, confirming each stSkillsJul 202675k182k
wonda-cliUsing the Wonda CLI to generate images, videos, music, and audio from the terminal — plus LinkedIn, Reddit, and X/Twitter research and automationSkillsJul 202666k137

co-design

222stars12forks

View on GitHub

How to install co-design?

am-will/swarms/co-design

$npx -y skills add am-will/swarms --skill co-design

Installs into the current project.

Summary

Only to be triggered by explicit /co-design commands. Parallel task executor that routes frontend/design tasks to Claude CLI print mode for superior design output.

Files · 1

View on GitHub

SKILL.md

1	# Co-Design: Parallel Task Executor with Design-Mode Routing
2
3	You are an Orchestrator for subagents. Use orchestration mode to parse plan files and delegate tasks to parallel subagents using task dependencies, in a loop, until all tasks are completed.
4
5	Key difference from standard parallel-task: When a task involves frontend design, styling, UI, or visual work, you launch it via `claude -p` (print mode) in a background shell instead of using the Task tool. This gives the design agent full CLI capabilities and tool access. All other tasks use normal Task tool subagents.
6
7	## Task Classification
8
9	Before launching each task, classify it as design or standard:
10
11	### Design Tasks (route to `claude -p`)
12	A task is a design task if it primarily involves ANY of:
13	- CSS, SCSS, Tailwind, or any styling
14	- HTML structure/templates for UI
15	- React/Vue/Svelte/Angular component markup and styling
16	- Layout, responsive design, grid, flexbox
17	- Design tokens, theme files, color schemes
18	- UI component creation (buttons, cards, modals, navbars, etc.)
19	- Animation, transitions, visual effects
20	- Accessibility (a11y) related to visual presentation
21	- Asset management (icons, images, fonts)
22	- Design system implementation
23
24	### Standard Tasks (route to Task tool subagent)
25	Everything else:
26	- Backend logic, API endpoints, database
27	- Business logic, services, utilities
28	- Configuration, tooling, CI/CD
29	- Testing (unless it's visual/snapshot testing)
30	- State management, data fetching
31	- Authentication, authorization
32
33	When in doubt: If a task mixes both (e.g., "create a form with validation"), route it as a design task since it has a UI component.
34
35	## Process
36
37	### Step 1: Parse Request
38
39	Extract from user request:
40	1. Plan file: The markdown plan to read
41	2. Task subset (optional): Specific task IDs to run
42
43	If no subset provided, run the full plan.
44
45	### Step 2: Read & Parse Plan
46
47	1. Find task subsections (e.g., `### T1:` or `### Task 1.1:`)
48	2. For each task, extract:
49	- Task ID and name
50	- depends_on list (from `- depends_on: [...]`)
51	- Full content (description, location, acceptance criteria, validation)
52	3. Build task list
53	4. Classify each task as `design` or `standard` based on the rules above
54	5. If a task subset was requested, filter the task list to only those IDs and their required dependencies.
55
56	### Step 3: Launch Tasks
57
58	For each unblocked task, launch it based on its classification:
59
60	#### For STANDARD tasks → Use Task tool subagent
61
62	Launch with:
63	- description: "Implement task [ID]: [name]"
64	- prompt: Use the Task Prompt Template below
65	- subagent_type: "parallel"
66
67	#### For DESIGN tasks → Use `claude -p` in background shell
68
69	Construct the exact same prompt you would give a subagent, then launch via Bash:
70
71	```bash
72	claude -p "YOUR_PROMPT_HERE" \
73	--allowedTools "Bash,Read,Edit,Write,Glob,Grep,WebFetch,WebSearch" \
74	--max-turns 50 \
75	--output-format text \
76	> /tmp/co-design-[TASK_ID]-output.log 2>&1 &
77	```
78
79	CRITICAL rules for design task launch:
80	1. The prompt content must be IDENTICAL to what you'd give a Task tool subagent (use the same Task Prompt Template)
81	2. Escape the prompt properly for shell execution - use a heredoc approach:
82	```bash
83	claude -p "$(cat <<'PROMPT_EOF'
84	[full prompt content here]
85	PROMPT_EOF
86	)" \
87	--allowedTools "Bash,Read,Edit,Write,Glob,Grep,WebFetch,WebSearch" \
88	--max-turns 50 \
89	--output-format text \
90	> /tmp/co-design-[TASK_ID]-output.log 2>&1 &
91	```
92	3. Log output to `/tmp/co-design-[TASK_ID]-output.log` for later inspection
93	4. Track the background PID so you can check completion
94	5. Launch ALL unblocked design tasks in parallel (multiple background shells)
95
96	Launch all unblocked tasks (both types) in parallel. A task is unblocked if all IDs in its depends_on list are complete.
97
98	### Task Prompt Template
99
100	Use this for BOTH standard subagents AND design-mode `claude -p` tasks:
101
102	```
103	You are implementing a specific task from a development plan.
104
105	## Context
106	- Plan: [filename]
107	- Goals: [relevant overview from plan]
108	- Dependencies: [prerequisites for this task]
109	- Related tasks: [tasks that depend on or are depended on by this task]
110	- Constraints: [risks from plan]
111
112	## Your Task
113	Task [ID]: [Name]
114
115	Location: [File paths]
116	Description: [Full description]
117
118	Acceptance Criteria:
119	[List from plan]
120
121	Validation:
122	[Tests or verification from plan]
123
124	## Instructions
125	1. Read the working plan and fully understand this task before coding.
126	2. Read all relevant files first, then do targeted codebase research (related modules, tests, call sites, and dependencies) to confirm the approach.
127	3. Default to TDD RED phase first using a `tdd_test_writer` subagent:
128	- Pass task context and acceptance criteria.
129	- Require tests-only edits.

People who liked this also viewed

6 picks

Type