model-deployment

Name: model-deployment
Author: awslabs

843stars132forks

How to install model-deployment?

awslabs/agent-plugins/model-deployment

$npx -y skills add awslabs/agent-plugins --skill model-deployment

Installs into the current project.

Summary

Generates code that deploys fine-tuned models from SageMaker Serverless Model Customization to SageMaker endpoints or Bedrock. Use when the user says "deploy my model", "create an endpoint", "make it available", or asks about deployment options. Identifies the correct deployment

Files · 1

View on GitHub

SKILL.md

1	# Model Deployment
2
3	Identifies the correct deployment pathway based on model characteristics and generates deployment code.
4
5	## Scope
6
7	This skill supports deploying Nova and OSS models that were fine-tuned through SageMaker Serverless Model Customization only.
8
9	Not supported:
10
11	- Base models (not fine-tuned)
12	- Models fine-tuned through other processes
13	- Full Fine-Tuning (FFT) — only LoRA fine-tuned models are supported
14
15	## Prerequisites
16
17	- The SDK environment has been verified (SDK version, region, execution role). If not done, activate the `sdk-getting-started` skill first.
18
19	---
20
21	## Principles
22
23	1. One thing at a time. Each response advances exactly one decision.
24	2. Confirm before proceeding. Wait for the user to agree before moving on. But don't re-ask questions already answered in the conversation — use what you know.
25	3. Don't read files until you need them. Only read pathway references after the pathway is confirmed.
26	4. Use what you know. If conversation history or artifacts already answer a question, confirm your understanding instead of asking again.
27
28	## Workflow
29
30	### Step 1: Identify the Training Job
31
32	You need the training job name or ARN. Check the conversation history first — the user may have already mentioned it, or it may be available from earlier steps in the workflow (e.g., fine-tuning). If not, ask the user.
33
34	Once you have the training job name or ARN, use the AWS MCP tool to look it up:
35
36	1. Use the AWS MCP tool `describe-training-job` and extract:
37	- S3 output path (from `ModelArtifacts.S3ModelArtifacts` or `OutputDataConfig.S3OutputPath`)
38	- IAM role ARN (from `RoleArn`)
39	- Region
40	2. Use the AWS MCP tool `list-tags` on the training job ARN and extract:
41	- Model ID from the `sagemaker-studio:jumpstart-model-id` tag
42	3. Determine the model type from the model ID:
43	- Contains "nova" (nova-micro, nova-lite, nova-pro) → Nova
44	- Llama, Mistral, Qwen, GPT-OSS, DeepSeek, etc. → OSS
45
46	Unsupported models: This skill only supports OSS and Nova models that were LoRA fine-tuned through SageMaker Serverless Model Customization. If the model doesn't match, tell the user this skill can't help and suggest the finetuning skill.
47
48	### Step 2: Determine Eligible Deployment Targets
49
50	Use the following table:
51
52	\| Model Type \| Eligible Targets \|
53	\| ---------- \| ------------------ \|
54	\| OSS \| SageMaker, Bedrock \|
55	\| Nova \| SageMaker, Bedrock \|
56
57	If only one target is eligible, confirm it with the user. Use details from Step 5.
58
59	If multiple targets are eligible, help the user decide. Use details from Step 5.
60
61	If no targets are eligible, tell the user and explain why.
62
63	### Step 3: Let the User Choose a Deployment Target
64
65	Present the eligible options to the user. Present these details to help them decide between SageMaker and Bedrock, if both are available options:
66
67	SageMaker Endpoint:
68
69	- Dedicated compute resources for consistent performance
70	- Control instance types and scaling
71	- Best for predictable workloads with specific latency requirements
72
73	Bedrock:
74
75	- Fully managed serverless inference
76	- Auto-scales instantly with no capacity planning
77	- Pay per request
78	- Best for variable workloads with fluctuating demand
79
80	Do NOT make a recommendation. Let the user choose.
81
82	Do NOT mention technical details like merged/unmerged weights, reference files, or APIs, unless the user asks.
83
84	⏸ Wait for user to select a deployment option.
85
86	### Step 4: Display License Agreement
87
88	Before proceeding to deployment, display the model's license or service terms to the user.
89
90	1. Read `references/model-licenses.md` and look up the model by its model ID (determined in Step 1).
91	2. Follow the instructions in the Notes column — use the exact phrasing provided.
92	3. If the model ID is not found in the table, warn the user that you could not find license information for their model and recommend they verify the license independently before proceeding.
93
94	⏸ Wait for the user to confirm before proceeding.
95
96	### Step 5: Follow Pathway Workflow
97
98	Read the reference file for the selected pathway and follow its instructions.
99
100	\| Model Type \| Deployment Target \| Reference \|
101	\| ---------- \| ----------------- \| ------------------------------------- \|
102	\| OSS \| SageMaker \| `references/deploy-oss-sagemaker.md` \|
103	\| OSS \| Bedrock \| `references/deploy-oss-bedrock.md` \|
104	\| Nova \| SageMaker \| `references/deploy-nova-sagemaker.md` \|
105	\| Nova \| Bedrock \| `references/deploy-nova-b

People who liked this also viewed

6 picks

Type

model-deployment

843stars132forks

View on GitHub

How to install model-deployment?

awslabs/agent-plugins/model-deployment

$npx -y skills add awslabs/agent-plugins --skill model-deployment

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Model Deployment
2
3	Identifies the correct deployment pathway based on model characteristics and generates deployment code.
4
5	## Scope
6
7	This skill supports deploying Nova and OSS models that were fine-tuned through SageMaker Serverless Model Customization only.
8
9	Not supported:
10
11	- Base models (not fine-tuned)
12	- Models fine-tuned through other processes
13	- Full Fine-Tuning (FFT) — only LoRA fine-tuned models are supported
14
15	## Prerequisites
16
17	- The SDK environment has been verified (SDK version, region, execution role). If not done, activate the `sdk-getting-started` skill first.
18
19	---
20
21	## Principles
22
23	1. One thing at a time. Each response advances exactly one decision.
24	2. Confirm before proceeding. Wait for the user to agree before moving on. But don't re-ask questions already answered in the conversation — use what you know.
25	3. Don't read files until you need them. Only read pathway references after the pathway is confirmed.
26	4. Use what you know. If conversation history or artifacts already answer a question, confirm your understanding instead of asking again.
27
28	## Workflow
29
30	### Step 1: Identify the Training Job
31
32	You need the training job name or ARN. Check the conversation history first — the user may have already mentioned it, or it may be available from earlier steps in the workflow (e.g., fine-tuning). If not, ask the user.
33
34	Once you have the training job name or ARN, use the AWS MCP tool to look it up:
35
36	1. Use the AWS MCP tool `describe-training-job` and extract:
37	- S3 output path (from `ModelArtifacts.S3ModelArtifacts` or `OutputDataConfig.S3OutputPath`)
38	- IAM role ARN (from `RoleArn`)
39	- Region
40	2. Use the AWS MCP tool `list-tags` on the training job ARN and extract:
41	- Model ID from the `sagemaker-studio:jumpstart-model-id` tag
42	3. Determine the model type from the model ID:
43	- Contains "nova" (nova-micro, nova-lite, nova-pro) → Nova
44	- Llama, Mistral, Qwen, GPT-OSS, DeepSeek, etc. → OSS
45
46	Unsupported models: This skill only supports OSS and Nova models that were LoRA fine-tuned through SageMaker Serverless Model Customization. If the model doesn't match, tell the user this skill can't help and suggest the finetuning skill.
47
48	### Step 2: Determine Eligible Deployment Targets
49
50	Use the following table:
51
52	\| Model Type \| Eligible Targets \|
53	\| ---------- \| ------------------ \|
54	\| OSS \| SageMaker, Bedrock \|
55	\| Nova \| SageMaker, Bedrock \|
56
57	If only one target is eligible, confirm it with the user. Use details from Step 5.
58
59	If multiple targets are eligible, help the user decide. Use details from Step 5.
60
61	If no targets are eligible, tell the user and explain why.
62
63	### Step 3: Let the User Choose a Deployment Target
64
65	Present the eligible options to the user. Present these details to help them decide between SageMaker and Bedrock, if both are available options:
66
67	SageMaker Endpoint:
68
69	- Dedicated compute resources for consistent performance
70	- Control instance types and scaling
71	- Best for predictable workloads with specific latency requirements
72
73	Bedrock:
74
75	- Fully managed serverless inference
76	- Auto-scales instantly with no capacity planning
77	- Pay per request
78	- Best for variable workloads with fluctuating demand
79
80	Do NOT make a recommendation. Let the user choose.
81
82	Do NOT mention technical details like merged/unmerged weights, reference files, or APIs, unless the user asks.
83
84	⏸ Wait for user to select a deployment option.
85
86	### Step 4: Display License Agreement
87
88	Before proceeding to deployment, display the model's license or service terms to the user.
89
90	1. Read `references/model-licenses.md` and look up the model by its model ID (determined in Step 1).
91	2. Follow the instructions in the Notes column — use the exact phrasing provided.
92	3. If the model ID is not found in the table, warn the user that you could not find license information for their model and recommend they verify the license independently before proceeding.
93
94	⏸ Wait for the user to confirm before proceeding.
95
96	### Step 5: Follow Pathway Workflow
97
98	Read the reference file for the selected pathway and follow its instructions.
99
100	\| Model Type \| Deployment Target \| Reference \|
101	\| ---------- \| ----------------- \| ------------------------------------- \|
102	\| OSS \| SageMaker \| `references/deploy-oss-sagemaker.md` \|
103	\| OSS \| Bedrock \| `references/deploy-oss-bedrock.md` \|
104	\| Nova \| SageMaker \| `references/deploy-nova-sagemaker.md` \|
105	\| Nova \| Bedrock \| `references/deploy-nova-b

People who liked this also viewed

6 picks

Type