agent-creation

Name: agent-creation
Author: DNYoussef

32stars6forks

How to install agent-creation?

dnyoussef/context-cascade/agent-creation

$npx -y skills add DNYoussef/context-cascade --skill agent-creation

Installs into the current project.

Summary

/*============================================================================*/ /* AGENT-CREATION SKILL :: VERILINGUA x VERIX EDITION */ /*============================================================================*/

Files · 1

View on GitHub

SKILL.md

1	/============================================================================/
2	/* AGENT-CREATION SKILL :: VERILINGUA x VERIX EDITION */
3	/============================================================================/
4
5	---
6	name: agent-creation
7	version: 1.0.0
8	description: \|
9	[assert\|neutral] Systematic agent creation using evidence-based prompting principles and 4-phase SOP methodology. Use when creating new specialist agents, refining existing agent prompts, or designing multi-agent syst [ground:given] [conf:0.95] [state:confirmed]
10	category: foundry
11	tags:
12	- foundry
13	- creation
14	- meta-tools
15	author: ruv
16	cognitive_frame:
17	primary: compositional
18	goal_analysis:
19	first_order: "Execute agent-creation workflow"
20	second_order: "Ensure quality and consistency"
21	third_order: "Enable systematic foundry processes"
22	---
23
24	/----------------------------------------------------------------------------/
25	/* S0 META-IDENTITY */
26	/----------------------------------------------------------------------------/
27
28	[define\|neutral] SKILL := {
29	name: "agent-creation",
30	category: "foundry",
31	version: "1.0.0",
32	layer: L1
33	} [ground:given] [conf:1.0] [state:confirmed]
34
35	/----------------------------------------------------------------------------/
36	/* S1 COGNITIVE FRAME */
37	/----------------------------------------------------------------------------/
38
39	[define\|neutral] COGNITIVE_FRAME := {
40	frame: "Compositional",
41	source: "German",
42	force: "Build from primitives?"
43	} [ground:cognitive-science] [conf:0.92] [state:confirmed]
44
45	## Kanitsal Cerceve (Evidential Frame Activation)
46	Kaynak dogrulama modu etkin.
47
48	/----------------------------------------------------------------------------/
49	/* S2 TRIGGER CONDITIONS */
50	/----------------------------------------------------------------------------/
51
52	[define\|neutral] TRIGGER_POSITIVE := {
53	keywords: ["agent-creation", "foundry", "workflow"],
54	context: "user needs agent-creation capability"
55	} [ground:given] [conf:1.0] [state:confirmed]
56
57	/----------------------------------------------------------------------------/
58	/* S3 CORE CONTENT */
59	/----------------------------------------------------------------------------/
60
61	<!-- SKILL SOP IMPROVEMENT v1.0 -->
62	## Skill Execution Criteria
63
64	### When to Use This Skill
65	- Creating new specialist agents with domain-specific expertise
66	- Refining existing agent system prompts for better performance
67	- Designing multi-agent coordination systems
68	- Implementing role-based agent hierarchies
69	- Building production-ready agents with embedded domain knowledge
70
71	### When NOT to Use This Skill
72	- For simple one-off tasks that don't need agent specialization
73	- When existing agents already cover the required domain
74	- For casual conversational interactions without systematic requirements
75	- When the task is better suited for a slash command or micro-skill
76
77	### Success Criteria
78	- [assert\|neutral] primary_outcome: "Production-ready agent with optimized system prompt, clear role definition, and validated performance" [ground:acceptance-criteria] [conf:0.90] [state:provisional]
79	- [assert\|neutral] quality_threshold: 0.9 [ground:acceptance-criteria] [conf:0.90] [state:provisional]
80	- [assert\|neutral] verification_method: "Agent successfully completes domain-specific tasks with consistent high-quality output, passes validation tests, and integrates with Claude Agent SDK" [ground:acceptance-criteria] [conf:0.90] [state:provisional]
81
82	### Edge Cases
83	- case: "Vague agent requirements"
84	handling: "Use Phase 1 (Initial Analysis) to research domain, identify patterns, and clarify scope before proceeding"
85	- case: "Overlapping agent capabilities"
86	handling: "Conduct agent registry search, identify gaps vs duplicates, propose consolidation or specialization"
87	- case: "Agent needs multiple conflicting personas"
88	handling: "Decompose into multiple focused agents with clear coordination pattern"
89
90	### Skill Guardrails
91	NEVER:
92	- "Create agents without deep domain research (skipping Phase 1 undermines quality)"
93	- "Use generic prompts without evidence-based techniques (CoT, few-shot, role-based)"
94	- "Skip validation testing (Phase 3) before considering agent production-ready"
95	- "Create agents that duplicate existing registry agents without justification"
96	ALWAYS:
97	- "Complete all 4 phases: Analysis -> Prompt Engineering -> Testing -> Integration"
98	- "Apply evidence-based prompting: Chain-of-Thought for reasoning, few-shot for patterns, clear role definition"
99	- "Validate with diverse test cases and measure against quality criteria"
100	- "Document agent capabilities, limitations, and integration points"
101
102	### Evidence-Based Execution
103	self_consistency: "After agent creation, test with same task multiple times to verify consistent o

People who liked this also viewed

6 picks

Type

agent-creation

32stars6forks

View on GitHub

How to install agent-creation?

dnyoussef/context-cascade/agent-creation

$npx -y skills add DNYoussef/context-cascade --skill agent-creation

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	/============================================================================/
2	/* AGENT-CREATION SKILL :: VERILINGUA x VERIX EDITION */
3	/============================================================================/
4
5	---
6	name: agent-creation
7	version: 1.0.0
8	description: \|
9	[assert\|neutral] Systematic agent creation using evidence-based prompting principles and 4-phase SOP methodology. Use when creating new specialist agents, refining existing agent prompts, or designing multi-agent syst [ground:given] [conf:0.95] [state:confirmed]
10	category: foundry
11	tags:
12	- foundry
13	- creation
14	- meta-tools
15	author: ruv
16	cognitive_frame:
17	primary: compositional
18	goal_analysis:
19	first_order: "Execute agent-creation workflow"
20	second_order: "Ensure quality and consistency"
21	third_order: "Enable systematic foundry processes"
22	---
23
24	/----------------------------------------------------------------------------/
25	/* S0 META-IDENTITY */
26	/----------------------------------------------------------------------------/
27
28	[define\|neutral] SKILL := {
29	name: "agent-creation",
30	category: "foundry",
31	version: "1.0.0",
32	layer: L1
33	} [ground:given] [conf:1.0] [state:confirmed]
34
35	/----------------------------------------------------------------------------/
36	/* S1 COGNITIVE FRAME */
37	/----------------------------------------------------------------------------/
38
39	[define\|neutral] COGNITIVE_FRAME := {
40	frame: "Compositional",
41	source: "German",
42	force: "Build from primitives?"
43	} [ground:cognitive-science] [conf:0.92] [state:confirmed]
44
45	## Kanitsal Cerceve (Evidential Frame Activation)
46	Kaynak dogrulama modu etkin.
47
48	/----------------------------------------------------------------------------/
49	/* S2 TRIGGER CONDITIONS */
50	/----------------------------------------------------------------------------/
51
52	[define\|neutral] TRIGGER_POSITIVE := {
53	keywords: ["agent-creation", "foundry", "workflow"],
54	context: "user needs agent-creation capability"
55	} [ground:given] [conf:1.0] [state:confirmed]
56
57	/----------------------------------------------------------------------------/
58	/* S3 CORE CONTENT */
59	/----------------------------------------------------------------------------/
60
61	<!-- SKILL SOP IMPROVEMENT v1.0 -->
62	## Skill Execution Criteria
63
64	### When to Use This Skill
65	- Creating new specialist agents with domain-specific expertise
66	- Refining existing agent system prompts for better performance
67	- Designing multi-agent coordination systems
68	- Implementing role-based agent hierarchies
69	- Building production-ready agents with embedded domain knowledge
70
71	### When NOT to Use This Skill
72	- For simple one-off tasks that don't need agent specialization
73	- When existing agents already cover the required domain
74	- For casual conversational interactions without systematic requirements
75	- When the task is better suited for a slash command or micro-skill
76
77	### Success Criteria
78	- [assert\|neutral] primary_outcome: "Production-ready agent with optimized system prompt, clear role definition, and validated performance" [ground:acceptance-criteria] [conf:0.90] [state:provisional]
79	- [assert\|neutral] quality_threshold: 0.9 [ground:acceptance-criteria] [conf:0.90] [state:provisional]
80	- [assert\|neutral] verification_method: "Agent successfully completes domain-specific tasks with consistent high-quality output, passes validation tests, and integrates with Claude Agent SDK" [ground:acceptance-criteria] [conf:0.90] [state:provisional]
81
82	### Edge Cases
83	- case: "Vague agent requirements"
84	handling: "Use Phase 1 (Initial Analysis) to research domain, identify patterns, and clarify scope before proceeding"
85	- case: "Overlapping agent capabilities"
86	handling: "Conduct agent registry search, identify gaps vs duplicates, propose consolidation or specialization"
87	- case: "Agent needs multiple conflicting personas"
88	handling: "Decompose into multiple focused agents with clear coordination pattern"
89
90	### Skill Guardrails
91	NEVER:
92	- "Create agents without deep domain research (skipping Phase 1 undermines quality)"
93	- "Use generic prompts without evidence-based techniques (CoT, few-shot, role-based)"
94	- "Skip validation testing (Phase 3) before considering agent production-ready"
95	- "Create agents that duplicate existing registry agents without justification"
96	ALWAYS:
97	- "Complete all 4 phases: Analysis -> Prompt Engineering -> Testing -> Integration"
98	- "Apply evidence-based prompting: Chain-of-Thought for reasoning, few-shot for patterns, clear role definition"
99	- "Validate with diverse test cases and measure against quality criteria"
100	- "Document agent capabilities, limitations, and integration points"
101
102	### Evidence-Based Execution
103	self_consistency: "After agent creation, test with same task multiple times to verify consistent o

People who liked this also viewed

6 picks

Type