make_plan-en

Name: make_plan-en
Author: agentscope-ai

24kstars2.8kforks

How to install make_plan-en?

agentscope-ai/qwenpaw/make_plan-en

$npx -y skills add agentscope-ai/QwenPaw --skill make_plan-en

Installs into the current project.

Summary

For external plan request scenarios, guides the Agent to request a clear, actionable, step-by-step plan from a stronger Agent via list_agents and chat_with_agent, emphasizing that the plan is executed by the requester, not by the consulted Agent.

Files · 1

View on GitHub

SKILL.md

1	# Make Plan
2
3	Use this Skill when you need to make an external plan request to a stronger Agent.
4
5	The goal of this Skill is not to outsource a task, but rather to:
6	- Request a plan from a stronger Agent
7	- The plan must consist of clear, actionable steps
8	- You execute the plan yourself
9	- Do not ask the consulted Agent to execute the task directly
10
11	How to invoke:
12	- Use `list_agents()` to check available Agents
13	- Use `chat_with_agent(...)` to request the stronger Agent to "make a plan"
14	- In `text`, explicitly include the prompt: only provide a step-by-step executable plan, do not execute
15	- To supplement or refine the original plan in a follow-up, pass `session_id`
16
17	Recommended invocation skeleton:
18
19	```text
20	list_agents()
21
22	chat_with_agent(
23	to_agent="<stronger_agent>",
24	text="[Agent <auto> requesting] Please help me create an execution plan for the following task. You do not need to execute the task -- just output clear, actionable steps in order.",
25	)
26	```
27
28	## Core Rules
29
30	This Skill handles one thing only:
31	1. Find a stronger Agent
32	2. Request that Agent to output an execution plan
33	3. Require the plan to be step-by-step, actionable, and in sequential order
34	4. You execute the plan yourself rather than asking the other Agent to do it for you
35
36	If what you truly need is a "plan", use this Skill.
37	If what you need is a final answer, architectural judgment, review conclusion, or direct execution, do not misuse this Skill.
38
39	## Applicable Scenarios
40
41	The following scenarios are suitable for making an external plan request:
42	- The task requires multi-step decomposition
43	- Steps have dependencies between them
44	- A clear sequence, checkpoints, or verification points are needed
45	- Multiple modules, files, systems, or roles are involved
46	- The user explicitly requests a plan before execution
47	- You want to obtain a more complete and reliable execution path first
48
49	## Do Not Use This Way
50
51	Do not use this Skill in the following situations:
52	- You actually want the other Agent to do the task for you
53	- What you really lack is a small piece of factual information, not a plan
54	- What you truly need is an architectural judgment or solution comparison
55	- You have not yet clearly understood the task objective
56
57	## What to Ask the Stronger Agent For
58
59	When calling `chat_with_agent(...)`, clearly state:
60	- You need a plan, not execution
61	- Steps must be concrete, not just abstract advice
62	- Preferably include verification methods and completion criteria
63
64	## Tool Invocation Rules
65
66	Execute in this order by default:
67	1. Call `list_agents()` to confirm available stronger Agents
68	2. Select the most suitable target Agent for creating the plan
69	3. Call `chat_with_agent(...)` to request the generation of an execution plan
70	4. After receiving the plan, execute it yourself -- do not continue outsourcing the task to the other Agent
71
72	When requesting the plan, explicitly state:
73	- Only the plan is needed, not execution
74	- Steps must be specific, not general advice
75	- Clearly request sequence, dependencies, checkpoints, and verification methods when needed
76
77	Common parameters for `chat_with_agent`:
78	- `to_agent`: Target Agent ID
79	- `text`: Request content; must explicitly state "only output the plan, do not execute the task"
80	- `session_id`: Optional; pass this to continue an existing conversation
81
82	Notes:
83	- `base_url` generally does not need to be provided; the tool will automatically resolve the current API address
84	- Not passing `session_id` will automatically create a new session
85
86	## Request Template
87
88	Please help me create an execution plan for the following task.
89	You do not need to execute the task itself -- just output the plan.
90
91	Task:
92	[What needs to be done]
93
94	Goal:
95	[What the final outcome should be]
96
97	Constraints:
98	- [...]
99	- [...]
100
101	Plan requirements:
102	1. Break it down into clear, executable steps
103	2. Indicate the recommended order
104	3. Point out dependencies, checkpoints, or verification methods where necessary
105	4. If there are obvious risks, include key points to watch out for
106
107	Output format:
108	[e.g., Output 5-8 numbered steps, 1-3 sentences each]
109
110	Example:
111
112	```text
113	chat_with_agent(
114	to_agent="strong_reasoner",
115	text="""
116	Please help me create an execution plan for the following task.
117	You do not need to execute the task itself -- just output the plan.
118
119	Task:
120	Modify a multi-module feature.
121
122	Goal:
123	Complete the change with low risk and avoid missing integration points.
124
125	Constraints:
126	- Minimize rework
127	- Include verifiable intermediate checkpoints
128
129	Plan requirements:
130	1. Break it down into actionable steps
131	2. Indicate the recommended order
132	3. Mark key dependencies and checkpoints
133	4. Include verification meth

People who liked this also viewed

6 picks

Type

make_plan-en

24kstars2.8kforks

View on GitHub

How to install make_plan-en?

agentscope-ai/qwenpaw/make_plan-en

$npx -y skills add agentscope-ai/QwenPaw --skill make_plan-en

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Make Plan
2
3	Use this Skill when you need to make an external plan request to a stronger Agent.
4
5	The goal of this Skill is not to outsource a task, but rather to:
6	- Request a plan from a stronger Agent
7	- The plan must consist of clear, actionable steps
8	- You execute the plan yourself
9	- Do not ask the consulted Agent to execute the task directly
10
11	How to invoke:
12	- Use `list_agents()` to check available Agents
13	- Use `chat_with_agent(...)` to request the stronger Agent to "make a plan"
14	- In `text`, explicitly include the prompt: only provide a step-by-step executable plan, do not execute
15	- To supplement or refine the original plan in a follow-up, pass `session_id`
16
17	Recommended invocation skeleton:
18
19	```text
20	list_agents()
21
22	chat_with_agent(
23	to_agent="<stronger_agent>",
24	text="[Agent <auto> requesting] Please help me create an execution plan for the following task. You do not need to execute the task -- just output clear, actionable steps in order.",
25	)
26	```
27
28	## Core Rules
29
30	This Skill handles one thing only:
31	1. Find a stronger Agent
32	2. Request that Agent to output an execution plan
33	3. Require the plan to be step-by-step, actionable, and in sequential order
34	4. You execute the plan yourself rather than asking the other Agent to do it for you
35
36	If what you truly need is a "plan", use this Skill.
37	If what you need is a final answer, architectural judgment, review conclusion, or direct execution, do not misuse this Skill.
38
39	## Applicable Scenarios
40
41	The following scenarios are suitable for making an external plan request:
42	- The task requires multi-step decomposition
43	- Steps have dependencies between them
44	- A clear sequence, checkpoints, or verification points are needed
45	- Multiple modules, files, systems, or roles are involved
46	- The user explicitly requests a plan before execution
47	- You want to obtain a more complete and reliable execution path first
48
49	## Do Not Use This Way
50
51	Do not use this Skill in the following situations:
52	- You actually want the other Agent to do the task for you
53	- What you really lack is a small piece of factual information, not a plan
54	- What you truly need is an architectural judgment or solution comparison
55	- You have not yet clearly understood the task objective
56
57	## What to Ask the Stronger Agent For
58
59	When calling `chat_with_agent(...)`, clearly state:
60	- You need a plan, not execution
61	- Steps must be concrete, not just abstract advice
62	- Preferably include verification methods and completion criteria
63
64	## Tool Invocation Rules
65
66	Execute in this order by default:
67	1. Call `list_agents()` to confirm available stronger Agents
68	2. Select the most suitable target Agent for creating the plan
69	3. Call `chat_with_agent(...)` to request the generation of an execution plan
70	4. After receiving the plan, execute it yourself -- do not continue outsourcing the task to the other Agent
71
72	When requesting the plan, explicitly state:
73	- Only the plan is needed, not execution
74	- Steps must be specific, not general advice
75	- Clearly request sequence, dependencies, checkpoints, and verification methods when needed
76
77	Common parameters for `chat_with_agent`:
78	- `to_agent`: Target Agent ID
79	- `text`: Request content; must explicitly state "only output the plan, do not execute the task"
80	- `session_id`: Optional; pass this to continue an existing conversation
81
82	Notes:
83	- `base_url` generally does not need to be provided; the tool will automatically resolve the current API address
84	- Not passing `session_id` will automatically create a new session
85
86	## Request Template
87
88	Please help me create an execution plan for the following task.
89	You do not need to execute the task itself -- just output the plan.
90
91	Task:
92	[What needs to be done]
93
94	Goal:
95	[What the final outcome should be]
96
97	Constraints:
98	- [...]
99	- [...]
100
101	Plan requirements:
102	1. Break it down into clear, executable steps
103	2. Indicate the recommended order
104	3. Point out dependencies, checkpoints, or verification methods where necessary
105	4. If there are obvious risks, include key points to watch out for
106
107	Output format:
108	[e.g., Output 5-8 numbered steps, 1-3 sentences each]
109
110	Example:
111
112	```text
113	chat_with_agent(
114	to_agent="strong_reasoner",
115	text="""
116	Please help me create an execution plan for the following task.
117	You do not need to execute the task itself -- just output the plan.
118
119	Task:
120	Modify a multi-module feature.
121
122	Goal:
123	Complete the change with low risk and avoid missing integration points.
124
125	Constraints:
126	- Minimize rework
127	- Include verifiable intermediate checkpoints
128
129	Plan requirements:
130	1. Break it down into actionable steps
131	2. Indicate the recommended order
132	3. Mark key dependencies and checkpoints
133	4. Include verification meth

People who liked this also viewed

6 picks

Type