conversation-to-skill

Name: conversation-to-skill
Author: Undertone0809

247stars30forks

How to install conversation-to-skill?

undertone0809/rudder/conversation-to-skill

$npx -y skills add Undertone0809/rudder --skill conversation-to-skill

Installs into the current project.

Summary

Turn the current conversation's workflow into a reusable agent skill. Use this whenever the user wants to make a workflow reusable, standardize a successful thread, package an agent capability, or convert an ad hoc process into a repeatable skill. Read the thread first, extract t

Files · 1

View on GitHub

SKILL.md

1	# Conversation To Skill
2
3	This skill turns the work happening in the current conversation into a reusable
4	agent skill.
5
6	Its job is not just to write `SKILL.md`.
7	Its job is to identify the durable workflow, separate it from one-off thread
8	noise, decide the right packaging and placement, and produce a skill that will
9	actually help a future agent perform better.
10
11	When useful, this skill should borrow the practical methods of `skill-creator`:
12	good descriptions, clean skill structure, eval-friendly organization, and an
13	improve-via-feedback loop. When this skill owns evaluation, bundle the relevant
14	toolchain locally under `agents/`, `assets/`, `eval-viewer/`, `scripts/`, and
15	`references/` so it stays self-contained instead of depending on another skill
16	directory at runtime.
17
18	## Use This Skill For
19
20	Use this skill when the user is trying to:
21
22	- turn the current task or workflow into a reusable skill
23	- capture a successful collaboration pattern for future runs
24	- standardize how a class of tasks should be handled
25	- extract a repeatable agent workflow from the current thread
26	- package a reasoning framework, execution sequence, or artifact pattern into a skill
27	- upgrade an existing draft skill so it is general, usable, and easier to trigger
28
29	Typical prompts:
30
31	- "Turn what we're doing into a skill."
32	- "I want this conversation to become an agent capability."
33	- "Make this reusable for next time."
34	- "Abstract this workflow into a Codex skill."
35	- "This should be a standard operating pattern, not a one-off chat."
36	- "Clean up this skill and make it actually reusable."
37
38	## Do Not Use This Skill For
39
40	Do not use this skill when the user mainly wants:
41
42	- a summary of the conversation without creating a reusable skill
43	- immediate execution of the task with no abstraction step
44	- a skill generated from multiple unseen threads you cannot inspect
45	- a rigid template that blindly copies file paths, project names, or temporary constraints
46	- a generic skill factory that ignores what was actually valuable in the conversation
47
48	If the conversation does not yet reveal a stable workflow, say that plainly and
49	help the user clarify the reusable part first.
50
51	## Core Principles
52
53	### Capture The Repeatable Value
54
55	The skill should capture the repeatable value, not the accidental details.
56
57	A good abstraction preserves:
58
59	- the job to be done
60	- the trigger conditions
61	- the critical inputs and outputs
62	- the sequence of reasoning or execution
63	- the judgment criteria that make the workflow valuable
64	- the boundaries and non-goals
65
66	A bad abstraction copies:
67
68	- temporary filenames
69	- irrelevant project-specific paths
70	- incidental tools that happened to be used once
71	- order-of-operations that are not actually essential
72	- user wording that does not generalize
73
74	### Explain Why, Not Just What
75
76	Prefer instructions that explain why a step matters.
77	Avoid brittle mandates unless the workflow truly requires them.
78
79	If you find yourself writing a long list of rigid commands with no reasoning,
80	you are probably transcribing the thread instead of building a skill.
81
82	### Choose The Smallest Useful Shape
83
84	Do not overbuild the skill.
85	Use the smallest structure that preserves the capability:
86
87	- `SKILL.md` only, when the workflow is mostly reasoning and sequencing
88	- `SKILL.md` plus `references/`, when the skill needs domain guidance
89	- `SKILL.md` plus `scripts/`, when deterministic repeated work should be bundled
90	- `SKILL.md` plus `evals/`, when the skill benefits from repeatable testing
91
92	### Decide Placement Before Writing Files
93
94	Pick the skill location before creating files so the paths stay stable:
95
96	- Global: `~/.agents/skills/<skill-name>`
97	- Project-based: `<project-path>/.agents/skills/<skill-name>`
98
99	If the user wants a global skill to be discoverable by Codex immediately, also
100	create:
101
102	- `~/.codex/skills/<skill-name>` as a symlink to the global skill directory
103
104	If you plan to run evals, place the workspace next to the skill directory as:
105
106	- `<skill-name>-workspace/`
107
108	## Default Workflow
109
110	Follow this sequence unless the user already provided enough structure.
111
112	### 1. Extract The Candidate Skill From The Current Thread
113
114	Read the current conversation first.
115	Pull out the real workflow before asking the user to restate everything.
116
117	Capture:
118
119	- what the user was trying to achieve
120	- what sequence of steps the agent followed or should follow
121	- which tools or artifacts mattered
122	- what correcti

People who liked this also viewed

6 picks

Type

conversation-to-skill

247stars30forks

View on GitHub

How to install conversation-to-skill?

undertone0809/rudder/conversation-to-skill

$npx -y skills add Undertone0809/rudder --skill conversation-to-skill

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Conversation To Skill
2
3	This skill turns the work happening in the current conversation into a reusable
4	agent skill.
5
6	Its job is not just to write `SKILL.md`.
7	Its job is to identify the durable workflow, separate it from one-off thread
8	noise, decide the right packaging and placement, and produce a skill that will
9	actually help a future agent perform better.
10
11	When useful, this skill should borrow the practical methods of `skill-creator`:
12	good descriptions, clean skill structure, eval-friendly organization, and an
13	improve-via-feedback loop. When this skill owns evaluation, bundle the relevant
14	toolchain locally under `agents/`, `assets/`, `eval-viewer/`, `scripts/`, and
15	`references/` so it stays self-contained instead of depending on another skill
16	directory at runtime.
17
18	## Use This Skill For
19
20	Use this skill when the user is trying to:
21
22	- turn the current task or workflow into a reusable skill
23	- capture a successful collaboration pattern for future runs
24	- standardize how a class of tasks should be handled
25	- extract a repeatable agent workflow from the current thread
26	- package a reasoning framework, execution sequence, or artifact pattern into a skill
27	- upgrade an existing draft skill so it is general, usable, and easier to trigger
28
29	Typical prompts:
30
31	- "Turn what we're doing into a skill."
32	- "I want this conversation to become an agent capability."
33	- "Make this reusable for next time."
34	- "Abstract this workflow into a Codex skill."
35	- "This should be a standard operating pattern, not a one-off chat."
36	- "Clean up this skill and make it actually reusable."
37
38	## Do Not Use This Skill For
39
40	Do not use this skill when the user mainly wants:
41
42	- a summary of the conversation without creating a reusable skill
43	- immediate execution of the task with no abstraction step
44	- a skill generated from multiple unseen threads you cannot inspect
45	- a rigid template that blindly copies file paths, project names, or temporary constraints
46	- a generic skill factory that ignores what was actually valuable in the conversation
47
48	If the conversation does not yet reveal a stable workflow, say that plainly and
49	help the user clarify the reusable part first.
50
51	## Core Principles
52
53	### Capture The Repeatable Value
54
55	The skill should capture the repeatable value, not the accidental details.
56
57	A good abstraction preserves:
58
59	- the job to be done
60	- the trigger conditions
61	- the critical inputs and outputs
62	- the sequence of reasoning or execution
63	- the judgment criteria that make the workflow valuable
64	- the boundaries and non-goals
65
66	A bad abstraction copies:
67
68	- temporary filenames
69	- irrelevant project-specific paths
70	- incidental tools that happened to be used once
71	- order-of-operations that are not actually essential
72	- user wording that does not generalize
73
74	### Explain Why, Not Just What
75
76	Prefer instructions that explain why a step matters.
77	Avoid brittle mandates unless the workflow truly requires them.
78
79	If you find yourself writing a long list of rigid commands with no reasoning,
80	you are probably transcribing the thread instead of building a skill.
81
82	### Choose The Smallest Useful Shape
83
84	Do not overbuild the skill.
85	Use the smallest structure that preserves the capability:
86
87	- `SKILL.md` only, when the workflow is mostly reasoning and sequencing
88	- `SKILL.md` plus `references/`, when the skill needs domain guidance
89	- `SKILL.md` plus `scripts/`, when deterministic repeated work should be bundled
90	- `SKILL.md` plus `evals/`, when the skill benefits from repeatable testing
91
92	### Decide Placement Before Writing Files
93
94	Pick the skill location before creating files so the paths stay stable:
95
96	- Global: `~/.agents/skills/<skill-name>`
97	- Project-based: `<project-path>/.agents/skills/<skill-name>`
98
99	If the user wants a global skill to be discoverable by Codex immediately, also
100	create:
101
102	- `~/.codex/skills/<skill-name>` as a symlink to the global skill directory
103
104	If you plan to run evals, place the workspace next to the skill directory as:
105
106	- `<skill-name>-workspace/`
107
108	## Default Workflow
109
110	Follow this sequence unless the user already provided enough structure.
111
112	### 1. Extract The Candidate Skill From The Current Thread
113
114	Read the current conversation first.
115	Pull out the real workflow before asking the user to restate everything.
116
117	Capture:
118
119	- what the user was trying to achieve
120	- what sequence of steps the agent followed or should follow
121	- which tools or artifacts mattered
122	- what correcti

People who liked this also viewed

6 picks

Type