orchestrator

Name: orchestrator
Author: KnoxOps

38stars4forks

How to install orchestrator?

knoxops/open-devops-skills/orchestrator

$npx -y skills add KnoxOps/open-devops-skills --skill orchestrator

Installs into the current project.

Summary

Cloud infrastructure cost optimization lifecycle -- identifies idle and underutilized resources, orchestrates safe isolation and observation, and manages the approve-decommission workflow to reduce cloud spend. Trigger phrases: "cost optimization", "infra cost", "zombie resources

Files · 1

View on GitHub

SKILL.md

1	## Input Parameters
2
3	\| Name \| Type \| Required \| Description \|
4	\|------\|------\|----------\|-------------\|
5	\| run_dir \| string \| Yes \| Workspace root directory \|
6	\| ssh_key_path \| string \| No \| SSH key path for cloud API calls \|
7	\| task_id \| string \| Yes \| Task ID for progress tracking \|
8
9	## Execution Flow
10
11	### Task Context
12
13	Before starting execution, initialize `task_context.json`:
14
15	```json
16	{
17	"task_id": "<task_id from input>",
18	"current_step": 0,
19	"current_step_id": null,
20	"status": "running",
21	"steps": {
22	"confirm_scope": "pending",
23	"confirm_connection": "pending",
24	"confirm_monitoring": "pending",
25	"generate_intent": "pending",
26	"generate_plan": "pending",
27	"collect_metrics": "pending",
28	"screen_resources": "pending",
29	"review_gate": "pending",
30	"deep_scan": "pending",
31	"phase_e_report": "pending",
32	"phase_e_select": "pending",
33	"phase_f_isolation": "pending",
34	"isolation_review_gate": "pending",
35	"execute_isolation": "pending",
36	"decision": "pending",
37	"phase_j_delete_review": "pending",
38	"final_report": "pending"
39	},
40	"updated_at": "<ISO timestamp>"
41	}
42	```
43
44	Update this file after each step completes. On error, set step status to `"failed"` and overall `status` to `"failed"`.
45
46	### Step 1: confirm_scope
47
48	Type: inline
49	Description: Confirm scan scope with user
50
51	## Execution
52	Follow these instructions:
53
54	Ask the user what to scan. One question at a time.
55	1. What environment? (prod/staging/dev/all)
56	2. Any resource types to include or exclude?
57	Prefer multiple choice. Skip if user already specified.
58
59
60	### Progress Tracking
61
62	After completing this step, update `task_context.json`:
63	- Set `current_step_id` to `"confirm_scope"`
64	- Set `steps.confirm_scope` to `"completed"`
65	### Step 2: confirm_connection
66
67	Type: inline
68	Description: Collect concrete connection details
69
70	## Execution
71	Follow the instructions in the prompt file `$PLUGINS/ico/skills/orchestrator/prompts/confirm-connection.prompt.md`.
72
73	Write the output to the specified output file.
74
75	## Output
76	- Schema: schemas/connection-config.schema.json
77	- File: connection_config.json
78
79	### Progress Tracking
80
81	After completing this step, update `task_context.json`:
82	- Set `current_step_id` to `"confirm_connection"`
83	- Set `steps.confirm_connection` to `"completed"`
84	### Step 3: confirm_monitoring
85
86	Type: inline
87	Description: Confirm monitoring and CI/CD data sources
88
89	## Execution
90	Follow these instructions:
91
92	Ask two questions:
93
94	1. "Do you have a monitoring system for historical metrics?"
95	Options: Prometheus, Datadog, CloudWatch, Azure Monitor, GCP Monitoring, or none.
96	If none — warn that only real-time snapshots will be used.
97
98	2. "Do you use a CI/CD or GitOps tool for deployments?"
99	Options: ArgoCD, FluxCD, Jenkins, GitLab CI, GitHub Actions, other, or none.
100	If they name one, ask how to access it (URL, token, namespace).
101
102	Record both in {run_dir}/connection_config.json under `monitoring` and `cicd`.
103
104
105	### Progress Tracking
106
107	After completing this step, update `task_context.json`:
108	- Set `current_step_id` to `"confirm_monitoring"`
109	- Set `steps.confirm_monitoring` to `"completed"`
110	### Step 4: generate_intent
111
112	Type: inline
113	Description: Generate intent_detection.json from user input
114
115	## Execution
116	Follow these instructions:
117
118	Write {run_dir}/intent_detection.json:
119	- entity_types: list of resource types to scan
120	- scope: {account, region, tag_filters}
121	- exclusions: list of excluded resource_ids or tags
122
123
124	Write the output to the specified output file.
125
126	## Output
127	- Schema: schemas/intent-detection.schema.json
128	- File: intent_detection.json
129
130	### Progress Tracking
131
132	After completing this step, update `task_context.json`:
133	- Set `current_step_id` to `"generate_intent"`
134	- Set `steps.generate_intent` to `"completed"`
135	### Step 5: generate_plan
136
137	Type: inline
138	Description: Dispatch Plan Agent with the orchestration template, present plan, wait for confirmation
139
140	## Execution
141	Follow these instructions:
142
143	1. Read the orchestration plan template: `$PLUGINS/ico/skills/orchestrator/prompts/orchestration-plan.agent.md`
144	2. Fill in any placeholders with info from intent_detection.json and user discussion
145	3. Use the Agent tool (subagent_type=general-purpose) with the filled template to produce an execution plan
146	4. Save the plan output verbatim to {run_dir}/execution_plan.md
147	5. Present the plan to user using the Write tool as markdown. Show: phase overview table, estimated duration, review checkpoints
148	6. WAIT for user confirmation before proceeding to execution.
149
150
151	### Progress Tracking
152
153	After completi

People who liked this also viewed

6 picks

Type

orchestrator

38stars4forks

View on GitHub

How to install orchestrator?

knoxops/open-devops-skills/orchestrator

$npx -y skills add KnoxOps/open-devops-skills --skill orchestrator

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	## Input Parameters
2
3	\| Name \| Type \| Required \| Description \|
4	\|------\|------\|----------\|-------------\|
5	\| run_dir \| string \| Yes \| Workspace root directory \|
6	\| ssh_key_path \| string \| No \| SSH key path for cloud API calls \|
7	\| task_id \| string \| Yes \| Task ID for progress tracking \|
8
9	## Execution Flow
10
11	### Task Context
12
13	Before starting execution, initialize `task_context.json`:
14
15	```json
16	{
17	"task_id": "<task_id from input>",
18	"current_step": 0,
19	"current_step_id": null,
20	"status": "running",
21	"steps": {
22	"confirm_scope": "pending",
23	"confirm_connection": "pending",
24	"confirm_monitoring": "pending",
25	"generate_intent": "pending",
26	"generate_plan": "pending",
27	"collect_metrics": "pending",
28	"screen_resources": "pending",
29	"review_gate": "pending",
30	"deep_scan": "pending",
31	"phase_e_report": "pending",
32	"phase_e_select": "pending",
33	"phase_f_isolation": "pending",
34	"isolation_review_gate": "pending",
35	"execute_isolation": "pending",
36	"decision": "pending",
37	"phase_j_delete_review": "pending",
38	"final_report": "pending"
39	},
40	"updated_at": "<ISO timestamp>"
41	}
42	```
43
44	Update this file after each step completes. On error, set step status to `"failed"` and overall `status` to `"failed"`.
45
46	### Step 1: confirm_scope
47
48	Type: inline
49	Description: Confirm scan scope with user
50
51	## Execution
52	Follow these instructions:
53
54	Ask the user what to scan. One question at a time.
55	1. What environment? (prod/staging/dev/all)
56	2. Any resource types to include or exclude?
57	Prefer multiple choice. Skip if user already specified.
58
59
60	### Progress Tracking
61
62	After completing this step, update `task_context.json`:
63	- Set `current_step_id` to `"confirm_scope"`
64	- Set `steps.confirm_scope` to `"completed"`
65	### Step 2: confirm_connection
66
67	Type: inline
68	Description: Collect concrete connection details
69
70	## Execution
71	Follow the instructions in the prompt file `$PLUGINS/ico/skills/orchestrator/prompts/confirm-connection.prompt.md`.
72
73	Write the output to the specified output file.
74
75	## Output
76	- Schema: schemas/connection-config.schema.json
77	- File: connection_config.json
78
79	### Progress Tracking
80
81	After completing this step, update `task_context.json`:
82	- Set `current_step_id` to `"confirm_connection"`
83	- Set `steps.confirm_connection` to `"completed"`
84	### Step 3: confirm_monitoring
85
86	Type: inline
87	Description: Confirm monitoring and CI/CD data sources
88
89	## Execution
90	Follow these instructions:
91
92	Ask two questions:
93
94	1. "Do you have a monitoring system for historical metrics?"
95	Options: Prometheus, Datadog, CloudWatch, Azure Monitor, GCP Monitoring, or none.
96	If none — warn that only real-time snapshots will be used.
97
98	2. "Do you use a CI/CD or GitOps tool for deployments?"
99	Options: ArgoCD, FluxCD, Jenkins, GitLab CI, GitHub Actions, other, or none.
100	If they name one, ask how to access it (URL, token, namespace).
101
102	Record both in {run_dir}/connection_config.json under `monitoring` and `cicd`.
103
104
105	### Progress Tracking
106
107	After completing this step, update `task_context.json`:
108	- Set `current_step_id` to `"confirm_monitoring"`
109	- Set `steps.confirm_monitoring` to `"completed"`
110	### Step 4: generate_intent
111
112	Type: inline
113	Description: Generate intent_detection.json from user input
114
115	## Execution
116	Follow these instructions:
117
118	Write {run_dir}/intent_detection.json:
119	- entity_types: list of resource types to scan
120	- scope: {account, region, tag_filters}
121	- exclusions: list of excluded resource_ids or tags
122
123
124	Write the output to the specified output file.
125
126	## Output
127	- Schema: schemas/intent-detection.schema.json
128	- File: intent_detection.json
129
130	### Progress Tracking
131
132	After completing this step, update `task_context.json`:
133	- Set `current_step_id` to `"generate_intent"`
134	- Set `steps.generate_intent` to `"completed"`
135	### Step 5: generate_plan
136
137	Type: inline
138	Description: Dispatch Plan Agent with the orchestration template, present plan, wait for confirmation
139
140	## Execution
141	Follow these instructions:
142
143	1. Read the orchestration plan template: `$PLUGINS/ico/skills/orchestrator/prompts/orchestration-plan.agent.md`
144	2. Fill in any placeholders with info from intent_detection.json and user discussion
145	3. Use the Agent tool (subagent_type=general-purpose) with the filled template to produce an execution plan
146	4. Save the plan output verbatim to {run_dir}/execution_plan.md
147	5. Present the plan to user using the Write tool as markdown. Show: phase overview table, estimated duration, review checkpoints
148	6. WAIT for user confirmation before proceeding to execution.
149
150
151	### Progress Tracking
152
153	After completi

People who liked this also viewed

6 picks

Type