council-builder

Name: council-builder
Author: berabuddies

555stars59forks

How to install council-builder?

berabuddies/semia/council-builder

$npx -y skills add berabuddies/Semia --skill council-builder

Installs into the current project.

Summary

Build a personalized team of AI agent personas for OpenClaw. Interviews the user, analyzes their workflow, then creates specialized agents with distinct personalities, adaptive model routing (Fast/Think/Deep/Strategic), weekly learning metrics, visual architecture docs, and inter

Files · 1

View on GitHub

SKILL.md

1	# Council Builder
2
3	Build a team of specialized AI agent personas tailored to the user's actual needs. Each agent gets a distinct personality, self-improvement capability, and clear coordination rules.
4
5	## Workflow
6
7	### Phase 1: Discovery
8
9	Interview the user to understand their world. Ask in batches of 2-3 questions max.
10
11	Round 1 - Identity:
12	- What do you do? (profession, main activities, industry)
13	- What tools and platforms do you use daily?
14
15	Round 2 - Pain Points:
16	- What tasks eat most of your time?
17	- Where do you feel you need the most help?
18
19	Round 3 - Preferences:
20	- What language(s) do you work in? (for agent communication style)
21	- Any specific domains you want covered? (coding, content, finance, research, scheduling, etc.)
22
23	Optional - History Analysis:
24	If the user has existing OpenClaw history, scan it for patterns:
25	- Check `memory/` files for recurring tasks
26	- Check existing workspace structure for active projects
27	- Check installed skills for current capabilities
28
29	Do NOT proceed to Phase 2 until confident you understand the user's needs. Ask follow-up questions if anything is unclear.
30
31	### Phase 2: Planning
32
33	Based on discovery, design the council:
34
35	1. Determine agent count: 3-7 agents. Fewer is better. Each agent must earn its existence.
36	2. Define each agent: Name, role, specialties, personality angle
37	3. Map coordination: Which agents feed data to which
38	4. Present the plan to the user in a clear table:
39
40	```
41	\| Agent \| Role \| Specialties \| Personality \|
42	\|-------\|------\|-------------\|-------------\|
43	\| [Name] \| [One-line role] \| [Key areas] \| [Personality angle] \|
44	```
45
46	5. Get explicit approval before building. Allow adjustments.
47
48	Naming agents:
49	- Give them memorable, short names (not generic like "Agent 1")
50	- Names should hint at their role but feel like characters
51	- Can be inspired by any theme the user likes, or choose strong standalone names
52	- See `references/example-councils.md` for naming patterns and complete council examples across different industries
53
54	### Phase 3: Building
55
56	Run the initialization script first to create the directory skeleton:
57	```bash
58	./scripts/init-council.sh <workspace-path> <agent-name-1> <agent-name-2> ...
59	```
60
61	Then, for each approved agent, populate the files. Read `references/soul-philosophy.md` before writing any SOUL.md.
62
63	Directory structure per agent:
64	```
65	agents/[agent-name]/
66	├── SOUL.md # Personality, role, rules (see soul-philosophy.md)
67	├── AGENTS.md # Agent-specific coordination rules
68	├── memory/ # Agent's memory directory
69	├── .learnings/ # Self-improvement logs
70	│ ├── LEARNINGS.md
71	│ ├── ERRORS.md
72	│ └── FEATURE_REQUESTS.md
73	└── [workspace dirs] # Role-specific output directories
74	```
75
76	For each agent's SOUL.md:
77	1. Read `references/soul-philosophy.md` for the writing guide
78	2. Read `assets/SOUL-TEMPLATE.md` for the structure
79	3. Customize deeply for this agent's role and personality
80	4. Every SOUL must be unique. No copy-paste between agents.
81
82	For each agent's AGENTS.md:
83	1. Use `assets/AGENT-AGENTS-TEMPLATE.md` as base
84	2. Define what this agent reads from and writes to
85	3. Define handoff rules with other agents
86
87	For .learnings/ files:
88	1. Copy structure from `assets/LEARNINGS-TEMPLATE.md`
89	2. Initialize empty log files
90
91	For the root AGENTS.md:
92	1. Use `assets/ROOT-AGENTS-TEMPLATE.md` as base
93	2. Create the routing table for all agents
94	3. Define file coordination map
95	4. Set up enforcement rules
96	5. Add adaptive model routing thresholds (Fast, Think, Deep, Strategic)
97
98	### Phase 4: Adaptive Routing Setup
99
100	Read `references/adaptive-routing.md`.
101
102	Set up an adaptive routing section in root AGENTS.md:
103	- Default to Fast
104	- Escalation thresholds for Think, Deep, Strategic
105	- De-escalation rule back to Fast after heavy reasoning
106	- High-tier model rate-limit fallback behavior
107
108	Also create visual architecture doc:
109	- `docs/architecture/ADAPTIVE-ROUTING-LEARNING.md` using `assets/ADAPTIVE-ROUTING-LEARNING-TEMPLATE.md`
110
111	### Phase 5: Self-Improvement Setup
112
113	Read `references/self-improvement.md` for the complete system.
114
115	Each agent gets built-in self-improvement:
116	- `.learnings/` directory with proper templates
117	- Detection triggers in SOUL.md

People who liked this also viewed

6 picks

Type

council-builder

555stars59forks

View on GitHub

How to install council-builder?

berabuddies/semia/council-builder

$npx -y skills add berabuddies/Semia --skill council-builder

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Council Builder
2
3	Build a team of specialized AI agent personas tailored to the user's actual needs. Each agent gets a distinct personality, self-improvement capability, and clear coordination rules.
4
5	## Workflow
6
7	### Phase 1: Discovery
8
9	Interview the user to understand their world. Ask in batches of 2-3 questions max.
10
11	Round 1 - Identity:
12	- What do you do? (profession, main activities, industry)
13	- What tools and platforms do you use daily?
14
15	Round 2 - Pain Points:
16	- What tasks eat most of your time?
17	- Where do you feel you need the most help?
18
19	Round 3 - Preferences:
20	- What language(s) do you work in? (for agent communication style)
21	- Any specific domains you want covered? (coding, content, finance, research, scheduling, etc.)
22
23	Optional - History Analysis:
24	If the user has existing OpenClaw history, scan it for patterns:
25	- Check `memory/` files for recurring tasks
26	- Check existing workspace structure for active projects
27	- Check installed skills for current capabilities
28
29	Do NOT proceed to Phase 2 until confident you understand the user's needs. Ask follow-up questions if anything is unclear.
30
31	### Phase 2: Planning
32
33	Based on discovery, design the council:
34
35	1. Determine agent count: 3-7 agents. Fewer is better. Each agent must earn its existence.
36	2. Define each agent: Name, role, specialties, personality angle
37	3. Map coordination: Which agents feed data to which
38	4. Present the plan to the user in a clear table:
39
40	```
41	\| Agent \| Role \| Specialties \| Personality \|
42	\|-------\|------\|-------------\|-------------\|
43	\| [Name] \| [One-line role] \| [Key areas] \| [Personality angle] \|
44	```
45
46	5. Get explicit approval before building. Allow adjustments.
47
48	Naming agents:
49	- Give them memorable, short names (not generic like "Agent 1")
50	- Names should hint at their role but feel like characters
51	- Can be inspired by any theme the user likes, or choose strong standalone names
52	- See `references/example-councils.md` for naming patterns and complete council examples across different industries
53
54	### Phase 3: Building
55
56	Run the initialization script first to create the directory skeleton:
57	```bash
58	./scripts/init-council.sh <workspace-path> <agent-name-1> <agent-name-2> ...
59	```
60
61	Then, for each approved agent, populate the files. Read `references/soul-philosophy.md` before writing any SOUL.md.
62
63	Directory structure per agent:
64	```
65	agents/[agent-name]/
66	├── SOUL.md # Personality, role, rules (see soul-philosophy.md)
67	├── AGENTS.md # Agent-specific coordination rules
68	├── memory/ # Agent's memory directory
69	├── .learnings/ # Self-improvement logs
70	│ ├── LEARNINGS.md
71	│ ├── ERRORS.md
72	│ └── FEATURE_REQUESTS.md
73	└── [workspace dirs] # Role-specific output directories
74	```
75
76	For each agent's SOUL.md:
77	1. Read `references/soul-philosophy.md` for the writing guide
78	2. Read `assets/SOUL-TEMPLATE.md` for the structure
79	3. Customize deeply for this agent's role and personality
80	4. Every SOUL must be unique. No copy-paste between agents.
81
82	For each agent's AGENTS.md:
83	1. Use `assets/AGENT-AGENTS-TEMPLATE.md` as base
84	2. Define what this agent reads from and writes to
85	3. Define handoff rules with other agents
86
87	For .learnings/ files:
88	1. Copy structure from `assets/LEARNINGS-TEMPLATE.md`
89	2. Initialize empty log files
90
91	For the root AGENTS.md:
92	1. Use `assets/ROOT-AGENTS-TEMPLATE.md` as base
93	2. Create the routing table for all agents
94	3. Define file coordination map
95	4. Set up enforcement rules
96	5. Add adaptive model routing thresholds (Fast, Think, Deep, Strategic)
97
98	### Phase 4: Adaptive Routing Setup
99
100	Read `references/adaptive-routing.md`.
101
102	Set up an adaptive routing section in root AGENTS.md:
103	- Default to Fast
104	- Escalation thresholds for Think, Deep, Strategic
105	- De-escalation rule back to Fast after heavy reasoning
106	- High-tier model rate-limit fallback behavior
107
108	Also create visual architecture doc:
109	- `docs/architecture/ADAPTIVE-ROUTING-LEARNING.md` using `assets/ADAPTIVE-ROUTING-LEARNING-TEMPLATE.md`
110
111	### Phase 5: Self-Improvement Setup
112
113	Read `references/self-improvement.md` for the complete system.
114
115	Each agent gets built-in self-improvement:
116	- `.learnings/` directory with proper templates
117	- Detection triggers in SOUL.md

People who liked this also viewed

6 picks

Type