agent-orchestration-improve-agent

Name: agent-orchestration-improve-agent
Author: SamarthaKV29

64stars12forks

How to install agent-orchestration-improve-agent?

samarthakv29/antigravity-god-mode/agent-orchestration-improve-agent

$npx -y skills add SamarthaKV29/antigravity-god-mode --skill agent-orchestration-improve-agent

Installs into the current project.

Summary

Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.

Files · 1

View on GitHub

SKILL.md

1	# Agent Performance Optimization Workflow
2
3	Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.
4
5	[Extended thinking: Agent optimization requires a data-driven approach combining performance metrics, user feedback analysis, and advanced prompt engineering techniques. Success depends on systematic evaluation, targeted improvements, and rigorous testing with rollback capabilities for production safety.]
6
7	## Use this skill when
8
9	- Improving an existing agent's performance or reliability
10	- Analyzing failure modes, prompt quality, or tool usage
11	- Running structured A/B tests or evaluation suites
12	- Designing iterative optimization workflows for agents
13
14	## Do not use this skill when
15
16	- You are building a brand-new agent from scratch
17	- There are no metrics, feedback, or test cases available
18	- The task is unrelated to agent performance or prompt quality
19
20	## Instructions
21
22	1. Establish baseline metrics and collect representative examples.
23	2. Identify failure modes and prioritize high-impact fixes.
24	3. Apply prompt and workflow improvements with measurable goals.
25	4. Validate with tests and roll out changes in controlled stages.
26
27	## Safety
28
29	- Avoid deploying prompt changes without regression testing.
30	- Roll back quickly if quality or safety metrics regress.
31
32	## Phase 1: Performance Analysis and Baseline Metrics
33
34	Comprehensive analysis of agent performance using context-manager for historical data collection.
35
36	### 1.1 Gather Performance Data
37
38	```
39	Use: context-manager
40	Command: analyze-agent-performance $ARGUMENTS --days 30
41	```
42
43	Collect metrics including:
44
45	- Task completion rate (successful vs failed tasks)
46	- Response accuracy and factual correctness
47	- Tool usage efficiency (correct tools, call frequency)
48	- Average response time and token consumption
49	- User satisfaction indicators (corrections, retries)
50	- Hallucination incidents and error patterns
51
52	### 1.2 User Feedback Pattern Analysis
53
54	Identify recurring patterns in user interactions:
55
56	- Correction patterns: Where users consistently modify outputs
57	- Clarification requests: Common areas of ambiguity
58	- Task abandonment: Points where users give up
59	- Follow-up questions: Indicators of incomplete responses
60	- Positive feedback: Successful patterns to preserve
61
62	### 1.3 Failure Mode Classification
63
64	Categorize failures by root cause:
65
66	- Instruction misunderstanding: Role or task confusion
67	- Output format errors: Structure or formatting issues
68	- Context loss: Long conversation degradation
69	- Tool misuse: Incorrect or inefficient tool selection
70	- Constraint violations: Safety or business rule breaches
71	- Edge case handling: Unusual input scenarios
72
73	### 1.4 Baseline Performance Report
74
75	Generate quantitative baseline metrics:
76
77	```
78	Performance Baseline:
79	- Task Success Rate: [X%]
80	- Average Corrections per Task: [Y]
81	- Tool Call Efficiency: [Z%]
82	- User Satisfaction Score: [1-10]
83	- Average Response Latency: [Xms]
84	- Token Efficiency Ratio: [X:Y]
85	```
86
87	## Phase 2: Prompt Engineering Improvements
88
89	Apply advanced prompt optimization techniques using prompt-engineer agent.
90
91	### 2.1 Chain-of-Thought Enhancement
92
93	Implement structured reasoning patterns:
94
95	```
96	Use: prompt-engineer
97	Technique: chain-of-thought-optimization
98	```
99
100	- Add explicit reasoning steps: "Let's approach this step-by-step..."
101	- Include self-verification checkpoints: "Before proceeding, verify that..."
102	- Implement recursive decomposition for complex tasks
103	- Add reasoning trace visibility for debugging
104
105	### 2.2 Few-Shot Example Optimization
106
107	Curate high-quality examples from successful interactions:
108
109	- Select diverse examples covering common use cases
110	- Include edge cases that previously failed
111	- Show both positive and negative examples with explanations
112	- Order examples from simple to complex
113	- Annotate examples with key decision points
114
115	Example structure:
116
117	```
118	Good Example:
119	Input: [User request]
120	Reasoning: [Step-by-step thought process]
121	Output: [Successful response]
122	Why this works: [Key success factors]
123
124	Bad Example:
125	Input: [Similar request]
126	Output: [Failed response]
127	Why this fails: [Specific issues]
128	Correct approach: [Fixed version]
129	```
130
131	### 2.3 Role Definition Refinement
132
133	Strengthen agent identity and capabilities:
134
135	- Core purpose: Clear, single-sentence mission
136	- Expertise domains: Specific knowledge areas
137	- Behavioral traits: Personality and interaction style
138	- Tool proficiency: Available tools and when to use them
139	- Constraints: What the agent should NOT do
140	- Success criteria: How to measure task completion
141
142	### 2.4 Constitutional AI Integration
143
144	Implement self-correction mechanisms:
145
146	```
147	Constitutional Principles:
148	1. Verify factual accuracy before responding
149	2. Self-check for potential biases or har

People who liked this also viewed

6 picks

Type

agent-orchestration-improve-agent

64stars12forks

View on GitHub

How to install agent-orchestration-improve-agent?

samarthakv29/antigravity-god-mode/agent-orchestration-improve-agent

$npx -y skills add SamarthaKV29/antigravity-god-mode --skill agent-orchestration-improve-agent

Installs into the current project.

Summary

Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.

Files · 1

View on GitHub

SKILL.md

1	# Agent Performance Optimization Workflow
2
3	Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.
4
5	[Extended thinking: Agent optimization requires a data-driven approach combining performance metrics, user feedback analysis, and advanced prompt engineering techniques. Success depends on systematic evaluation, targeted improvements, and rigorous testing with rollback capabilities for production safety.]
6
7	## Use this skill when
8
9	- Improving an existing agent's performance or reliability
10	- Analyzing failure modes, prompt quality, or tool usage
11	- Running structured A/B tests or evaluation suites
12	- Designing iterative optimization workflows for agents
13
14	## Do not use this skill when
15
16	- You are building a brand-new agent from scratch
17	- There are no metrics, feedback, or test cases available
18	- The task is unrelated to agent performance or prompt quality
19
20	## Instructions
21
22	1. Establish baseline metrics and collect representative examples.
23	2. Identify failure modes and prioritize high-impact fixes.
24	3. Apply prompt and workflow improvements with measurable goals.
25	4. Validate with tests and roll out changes in controlled stages.
26
27	## Safety
28
29	- Avoid deploying prompt changes without regression testing.
30	- Roll back quickly if quality or safety metrics regress.
31
32	## Phase 1: Performance Analysis and Baseline Metrics
33
34	Comprehensive analysis of agent performance using context-manager for historical data collection.
35
36	### 1.1 Gather Performance Data
37
38	```
39	Use: context-manager
40	Command: analyze-agent-performance $ARGUMENTS --days 30
41	```
42
43	Collect metrics including:
44
45	- Task completion rate (successful vs failed tasks)
46	- Response accuracy and factual correctness
47	- Tool usage efficiency (correct tools, call frequency)
48	- Average response time and token consumption
49	- User satisfaction indicators (corrections, retries)
50	- Hallucination incidents and error patterns
51
52	### 1.2 User Feedback Pattern Analysis
53
54	Identify recurring patterns in user interactions:
55
56	- Correction patterns: Where users consistently modify outputs
57	- Clarification requests: Common areas of ambiguity
58	- Task abandonment: Points where users give up
59	- Follow-up questions: Indicators of incomplete responses
60	- Positive feedback: Successful patterns to preserve
61
62	### 1.3 Failure Mode Classification
63
64	Categorize failures by root cause:
65
66	- Instruction misunderstanding: Role or task confusion
67	- Output format errors: Structure or formatting issues
68	- Context loss: Long conversation degradation
69	- Tool misuse: Incorrect or inefficient tool selection
70	- Constraint violations: Safety or business rule breaches
71	- Edge case handling: Unusual input scenarios
72
73	### 1.4 Baseline Performance Report
74
75	Generate quantitative baseline metrics:
76
77	```
78	Performance Baseline:
79	- Task Success Rate: [X%]
80	- Average Corrections per Task: [Y]
81	- Tool Call Efficiency: [Z%]
82	- User Satisfaction Score: [1-10]
83	- Average Response Latency: [Xms]
84	- Token Efficiency Ratio: [X:Y]
85	```
86
87	## Phase 2: Prompt Engineering Improvements
88
89	Apply advanced prompt optimization techniques using prompt-engineer agent.
90
91	### 2.1 Chain-of-Thought Enhancement
92
93	Implement structured reasoning patterns:
94
95	```
96	Use: prompt-engineer
97	Technique: chain-of-thought-optimization
98	```
99
100	- Add explicit reasoning steps: "Let's approach this step-by-step..."
101	- Include self-verification checkpoints: "Before proceeding, verify that..."
102	- Implement recursive decomposition for complex tasks
103	- Add reasoning trace visibility for debugging
104
105	### 2.2 Few-Shot Example Optimization
106
107	Curate high-quality examples from successful interactions:
108
109	- Select diverse examples covering common use cases
110	- Include edge cases that previously failed
111	- Show both positive and negative examples with explanations
112	- Order examples from simple to complex
113	- Annotate examples with key decision points
114
115	Example structure:
116
117	```
118	Good Example:
119	Input: [User request]
120	Reasoning: [Step-by-step thought process]
121	Output: [Successful response]
122	Why this works: [Key success factors]
123
124	Bad Example:
125	Input: [Similar request]
126	Output: [Failed response]
127	Why this fails: [Specific issues]
128	Correct approach: [Fixed version]
129	```
130
131	### 2.3 Role Definition Refinement
132
133	Strengthen agent identity and capabilities:
134
135	- Core purpose: Clear, single-sentence mission
136	- Expertise domains: Specific knowledge areas
137	- Behavioral traits: Personality and interaction style
138	- Tool proficiency: Available tools and when to use them
139	- Constraints: What the agent should NOT do
140	- Success criteria: How to measure task completion
141
142	### 2.4 Constitutional AI Integration
143
144	Implement self-correction mechanisms:
145
146	```
147	Constitutional Principles:
148	1. Verify factual accuracy before responding
149	2. Self-check for potential biases or har

People who liked this also viewed

6 picks

Type