Skip to content

Commit 799e319

Browse files
badhopebadhope
committed
feat: Add 15 comprehensive workflow skills from task simulation
1 parent 1437af2 commit 799e319

File tree

16 files changed

+843
-5
lines changed

16 files changed

+843
-5
lines changed

‎.trae/skills/workflows/ab-testing/SKILL.md‎

Lines changed: 55 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,55 @@
1+
---
2+
name: ab-testing
3+
description: "Tier 2: A/B testing and experimentation. Design experiments, analyze results, statistical significance. Keywords: A/B testing, experimentation, hypothesis testing, A/B测试, 实验设计"
4+
layer: workflow
5+
role: data-scientist
6+
tier: 2
7+
version: 5.0.0
8+
architecture: handoff-chain
9+
invokes:
10+
- human-in-the-loop
11+
- iteration-controller
12+
tags:
13+
- experimentation
14+
- ab-testing
15+
- statistics
16+
- optimization
17+
---
18+
19+
# A/B Testing
20+
21+
## Overview
22+
23+
A/B testing and experimentation framework.
24+
25+
## Key Capabilities
26+
27+
- Experiment design
28+
- Hypothesis formulation
29+
- Sample size calculation
30+
- Statistical analysis
31+
- Result interpretation
32+
- Recommendation
33+
34+
## Statistical Methods
35+
36+
- Hypothesis testing (t-test, chi-square)
37+
- p-value calculation
38+
- Confidence intervals
39+
- Bayesian A/B testing
40+
- Multi-armed bandits
41+
42+
## Process Flow
43+
44+
1. **Hypothesize** - Formulate hypothesis
45+
2. **Design** - Design experiment
46+
3. **Run** - Run experiment
47+
4. **Analyze** - Statistical analysis
48+
5. **Decide** - Interpret and decide
49+
50+
## Output Artifacts
51+
52+
- Experiment design document
53+
- Statistical analysis report
54+
- Result interpretation
55+
- Actionable recommendations

‎.trae/skills/workflows/conversation-design/SKILL.md‎

Lines changed: 56 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,56 @@
1+
---
2+
name: conversation-design
3+
description: "Tier 2: Conversation flow design for chatbots. Dialog state management, intent design, user journey. Keywords: conversation design, chatbot, dialog flow, 对话设计, 对话流程"
4+
layer: workflow
5+
role: conversation-designer
6+
tier: 2
7+
version: 5.0.0
8+
architecture: handoff-chain
9+
invokes:
10+
- agent-development
11+
- prompt-engineering
12+
tags:
13+
- chatbot
14+
- conversation
15+
- design
16+
- dialog
17+
---
18+
19+
# Conversation Design
20+
21+
## Overview
22+
23+
Conversation flow design for chatbots and voice assistants.
24+
25+
## Key Capabilities
26+
27+
- Dialog state management
28+
- Intent and entity design
29+
- Conversation flow mapping
30+
- Fallback handling
31+
- Personality and tone design
32+
- Multi-turn conversation design
33+
34+
## Design Methodologies
35+
36+
- Voice User Interface (VUI) design
37+
- Conversation patterns
38+
- User-centered design
39+
- Progressive disclosure
40+
- Error recovery strategies
41+
42+
## Process Flow
43+
44+
1. **Map** - User journey mapping
45+
2. **Design** - Conversation flow design
46+
3. **Define** - Intents and entities
47+
4. **Write** - Dialog scripts
48+
5. **Test** - Conversation testing
49+
50+
## Output Artifacts
51+
52+
- Conversation flow diagram
53+
- Intent definition document
54+
- Dialog scripts
55+
- State machine definition
56+
- Testing scenarios

‎.trae/skills/workflows/data-augmentation/SKILL.md‎

Lines changed: 53 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,53 @@
1+
---
2+
name: data-augmentation
3+
description: "Tier 2: Data augmentation for training data. Generate synthetic data, expand datasets, improve model robustness. Keywords: data augmentation, synthetic data, dataset expansion, 数据增强, 合成数据"
4+
layer: workflow
5+
role: data-engineer
6+
tier: 2
7+
version: 5.0.0
8+
architecture: handoff-chain
9+
invokes:
10+
- data-validation
11+
- etl
12+
tags:
13+
- data
14+
- augmentation
15+
- synthetic-data
16+
- machine-learning
17+
---
18+
19+
# Data Augmentation
20+
21+
## Overview
22+
23+
Data augmentation and synthetic data generation.
24+
25+
## Key Capabilities
26+
27+
- Text data augmentation
28+
- Synthetic data generation
29+
- Dataset expansion
30+
- Data balancing
31+
- Noise injection
32+
- Data quality improvement
33+
34+
## Augmentation Techniques
35+
36+
- Text: paraphrasing, back-translation, synonym replacement
37+
- Tabular: SMOTE, noise injection, synthetic sampling
38+
- General: GANs, VAEs, LLM-based generation
39+
40+
## Process Flow
41+
42+
1. **Analyze** - Analyze dataset characteristics
43+
2. **Select** - Select augmentation techniques
44+
3. **Generate** - Generate augmented data
45+
4. **Validate** - Validate data quality
46+
5. **Merge** - Merge with original dataset
47+
48+
## Output Artifacts
49+
50+
- Augmented dataset
51+
- Data augmentation report
52+
- Quality validation results
53+
- Dataset statistics

‎.trae/skills/workflows/database-design/SKILL.md‎

Lines changed: 55 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,55 @@
1+
---
2+
name: database-design
3+
description: "Tier 2: Database design and modeling. ER diagrams, schema design, normalization, optimization. Keywords: database design, schema, ER diagram, normalization, 数据库设计, 数据建模"
4+
layer: workflow
5+
role: database-architect
6+
tier: 2
7+
version: 5.0.0
8+
architecture: handoff-chain
9+
invokes:
10+
- sql-optimization
11+
- mongodb
12+
tags:
13+
- database
14+
- design
15+
- modeling
16+
- schema
17+
---
18+
19+
# Database Design
20+
21+
## Overview
22+
23+
Database design and data modeling.
24+
25+
## Key Capabilities
26+
27+
- Entity-Relationship (ER) modeling
28+
- Schema design
29+
- Normalization (1NF to 5NF)
30+
- Denormalization for performance
31+
- Indexing strategy
32+
- Query optimization
33+
34+
## Supported Databases
35+
36+
- Relational: PostgreSQL, MySQL, SQL Server
37+
- NoSQL: MongoDB, Redis, Cassandra
38+
- NewSQL: CockroachDB, TiDB
39+
- Graph: Neo4j
40+
41+
## Process Flow
42+
43+
1. **Model** - Create ER diagram
44+
2. **Normalize** - Apply normalization rules
45+
3. **Optimize** - Indexing and query optimization
46+
4. **Schema** - Generate schema DDL
47+
5. **Document** - Database documentation
48+
49+
## Output Artifacts
50+
51+
- ER diagram
52+
- Database schema
53+
- DDL scripts
54+
- Indexing strategy
55+
- Query optimization guide

‎.trae/skills/workflows/launch-checklist/SKILL.md‎

Lines changed: 58 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,58 @@
1+
---
2+
name: launch-checklist
3+
description: "Tier 2: Launch checklist and go-live preparation. Pre-launch checks, deployment verification, rollback plan. Keywords: launch checklist, go-live, deployment, rollback, 上线检查清单, 发布准备"
4+
layer: workflow
5+
role: release-manager
6+
tier: 2
7+
version: 5.0.0
8+
architecture: handoff-chain
9+
invokes:
10+
- git-operations
11+
- ci-cd-pipeline
12+
- fallback-manager
13+
tags:
14+
- launch
15+
- deployment
16+
- release
17+
- checklist
18+
---
19+
20+
# Launch Checklist
21+
22+
## Overview
23+
24+
Go-live checklist and release preparation.
25+
26+
## Key Capabilities
27+
28+
- Pre-launch checklist
29+
- Deployment verification
30+
- Rollback plan
31+
- Post-launch monitoring
32+
- Incident response
33+
34+
## Checklist Categories
35+
36+
- Code and testing
37+
- Infrastructure
38+
- Security
39+
- Performance
40+
- Monitoring
41+
- Documentation
42+
- Communication
43+
44+
## Process Flow
45+
46+
1. **Prepare** - Review and execute checklist
47+
2. **Verify** - Deployment verification
48+
3. **Monitor** - Post-launch monitoring
49+
4. **Respond** - Incident response if needed
50+
5. **Document** - Post-launch review
51+
52+
## Output Artifacts
53+
54+
- Launch checklist report
55+
- Deployment verification
56+
- Rollback plan
57+
- Post-launch monitoring plan
58+
- Post-mortem template

‎.trae/skills/workflows/model-evaluator/SKILL.md‎

Lines changed: 55 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,55 @@
1+
---
2+
name: model-evaluator
3+
description: "Tier 2: AI model evaluation and selection. Compare models, benchmark performance, select optimal model. Keywords: model evaluation, benchmark, model selection, LLM evaluation, 模型评估, 模型选择"
4+
layer: workflow
5+
role: ml-engineer
6+
tier: 2
7+
version: 5.0.0
8+
architecture: handoff-chain
9+
invokes:
10+
- llm-evaluation
11+
- research-workflow
12+
tags:
13+
- ai
14+
- llm
15+
- evaluation
16+
- benchmarking
17+
---
18+
19+
# Model Evaluator
20+
21+
## Overview
22+
23+
AI model evaluation and selection.
24+
25+
## Key Capabilities
26+
27+
- Model benchmarking
28+
- Performance comparison
29+
- Cost analysis
30+
- Latency testing
31+
- Quality assessment
32+
- Model selection recommendation
33+
34+
## Evaluation Dimensions
35+
36+
- Quality (accuracy, relevance, coherence)
37+
- Performance (latency, throughput)
38+
- Cost (token usage, API cost)
39+
- Safety (harm avoidance, bias)
40+
- Capabilities (reasoning, coding, creativity)
41+
42+
## Process Flow
43+
44+
1. **Define** - Define evaluation criteria
45+
2. **Benchmark** - Run benchmarks
46+
3. **Compare** - Compare model performance
47+
4. **Analyze** - Cost-benefit analysis
48+
5. **Recommend** - Model selection recommendation
49+
50+
## Output Artifacts
51+
52+
- Model evaluation report
53+
- Performance comparison matrix
54+
- Cost analysis
55+
- Recommendation document

‎.trae/skills/workflows/model-finetuning/SKILL.md‎

Lines changed: 55 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,55 @@
1+
---
2+
name: model-finetuning
3+
description: "Tier 2: Model fine-tuning and customization. Fine-tune LLMs on custom data, LoRA, QLoRA. Keywords: model fine-tuning, LoRA, QLoRA, customization, 模型微调, 定制化"
4+
layer: workflow
5+
role: ml-engineer
6+
tier: 2
7+
version: 5.0.0
8+
architecture: handoff-chain
9+
invokes:
10+
- llm-evaluation
11+
- data-augmentation
12+
tags:
13+
- ai
14+
- llm
15+
- fine-tuning
16+
- customization
17+
---
18+
19+
# Model Fine-tuning
20+
21+
## Overview
22+
23+
Model fine-tuning and customization.
24+
25+
## Key Capabilities
26+
27+
- LoRA (Low-Rank Adaptation)
28+
- QLoRA (Quantized LoRA)
29+
- Full fine-tuning
30+
- Instruction tuning
31+
- Alignment tuning
32+
- Evaluation before/after
33+
34+
## Supported Frameworks
35+
36+
- Hugging Face Transformers
37+
- PEFT (Parameter-Efficient Fine-Tuning)
38+
- LoRAX
39+
- Axolotl
40+
- LLaMA Factory
41+
42+
## Process Flow
43+
44+
1. **Prepare** - Prepare training data
45+
2. **Configure** - Configure fine-tuning parameters
46+
3. **Train** - Execute fine-tuning
47+
4. **Evaluate** - Evaluate fine-tuned model
48+
5. **Deploy** - Deploy fine-tuned model
49+
50+
## Output Artifacts
51+
52+
- Fine-tuned model weights
53+
- Training logs
54+
- Evaluation report
55+
- Deployment guide

0 commit comments

Comments
 (0)