Prompt Polisher
Transform messy, unstructured input into polished, Claude-optimized prompts using Anthropic's best practices.
When to Activate
Activate this skill when:
- User provides voice transcription content (contains filler words, self-corrections, casual speech)
- User dumps stream-of-consciousness notes
- User pastes rough document content that needs structuring
- User explicitly asks to "polish", "clean up", or "turn this into a prompt"
- Input is clearly unstructured and intended as task instructions
Workflow
Execute these stages in order:
Stage 1: Voice Cleanup
If the input appears to be voice transcription or casual speech, clean it up:
Remove:
- Filler words: "um", "uh", "like", "you know", "basically", "actually", "literally", "right?"
- Self-corrections: "no wait", "I mean", "actually scratch that", "let me rephrase"
- False starts: repeated or abandoned sentence beginnings
- Verbal tics: "okay so", "let me think", "hmm"
Normalize:
- Convert spoken patterns to written form
- Fix run-on sentences
- Add proper punctuation
Stage 2: Intent Extraction
Parse the cleaned input to identify:
| Element |
What to Find |
| Core task |
What does the user actually want done? |
| Referenced files |
Documents, project files, URLs mentioned |
| Constraints |
Limitations, requirements, boundaries |
| Preferences |
Style, format, approach preferences |
| Success criteria |
What does "done" look like? |
| Scope |
Single task vs. multi-step project |
Stage 3: Gap Detection
Scan for missing or ambiguous elements:
If gaps exist, ask ALL clarifying questions at once:
Before I polish this prompt, I need to clarify a few things:
1. [First gap-specific question]
2. [Second gap-specific question]
3. Which Claude model are you using? (Opus 4.5 / Sonnet 4.5 / Not sure)
Wait for user response before proceeding.
Stage 4: Apply Best Practices
Reference reference.md for detailed guidelines. Apply these transformations:
Universal (Claude 4.x):
- Make instructions explicit and direct (no hints)
- Add contextual framing (explain WHY, not just WHAT)
- Use positive framing ("use prose" not "don't use bullets")
- Structure with XML tags for complex inputs
- Use action-oriented language ("Change this" not "Can you suggest")
- Include chain of thought prompts for complex reasoning
- Define clear success criteria
- Use normal tone (avoid aggressive CAPS/MUST/CRITICAL)
If Opus 4.5:
- Add anti-over-engineering guardrail: "Keep solutions simple. Only make changes directly requested."
- Avoid the word "think" → use "consider", "evaluate", "believe"
- Leverage planning strength: "Build an editable plan before executing"
- Keep prompts token-efficient (don't over-pad)
- Provide policy/constraint context when relevant
If Sonnet 4.5:
- Break complex tasks into explicit steps
- Optimize for speed/throughput
- Add more explicit step-by-step guidance for intricate tasks
Stage 5: Structure the Prompt
Generate the polished prompt using this template:
<context>
[Background info, relevant project context, referenced files, WHY this task matters]
</context>
<task>
[Clear, explicit statement of what needs to be done - action-oriented]
</task>
<requirements>
[Specific constraints, boundaries, must-haves, technical choices]
</requirements>
<success_criteria>
[What "done" looks like, how to verify completion]
</success_criteria>
<approach> <!-- Optional: include for complex tasks -->
[Suggested approach OR "Build an editable plan before executing"]
</approach>
Formatting rules:
- Keep each section focused and concise
- Use bullet points only when listing multiple items
- Omit sections that aren't relevant to the task
- For simple tasks, a streamlined version without XML tags is acceptable
Stage 6: Preview for Approval
Present the polished prompt to the user:
Here's your polished prompt:
---
[POLISHED PROMPT CONTENT]
---
**Options:**
- **Execute** - Run this prompt now
- **Edit** - I'll modify something first
- **Redo** - Start over with different clarifications
Wait for user approval before executing.
Stage 7: Execute or Iterate
Based on user choice:
- Execute: Immediately act on the polished prompt as if the user had typed it directly
- Edit: User modifies the prompt, then re-preview
- Redo: Return to Stage 1 with new input or Stage 3 for new clarifications
Important Notes
- Always preview before executing - never skip the approval step
- If the input is already well-structured, acknowledge this and offer minor polish only
- Learn from user feedback - if they frequently edit a certain way, note the pattern
- For very short/simple inputs, use a streamlined format without full XML structure
- The goal is USEFUL prompts, not LONG prompts - be concise
See Also
- reference.md - Detailed Anthropic best practices for Claude 4.x, Opus 4.5, and Sonnet 4.5
- examples.md - Before/after transformation examples
1---
2name: prompt-polisher
3description: Use when receiving messy, unstructured input like voice transcriptions, stream-of-consciousness notes, or rough document content that needs to be transformed into a polished, optimized prompt. Cleans up filler words, extracts intent, asks clarifying questions, applies Claude 4.x/Opus 4.5/Sonnet 4.5 best practices, and previews the polished prompt for approval before execution. Trigger phrases include "polish this", "clean this up", "turn this into a prompt", or when input is clearly rough/unstructured.
4metadata:
5 source: https://skillmd.com/skills/cnfjlhj/prompt-polisher
6---
7
8# Prompt Polisher
9
10Transform messy, unstructured input into polished, Claude-optimized prompts using Anthropic's best practices.
11
12## When to Activate
13
14Activate this skill when:
15- User provides voice transcription content (contains filler words, self-corrections, casual speech)
16- User dumps stream-of-consciousness notes
17- User pastes rough document content that needs structuring
18- User explicitly asks to "polish", "clean up", or "turn this into a prompt"
19- Input is clearly unstructured and intended as task instructions
20
21## Workflow
22
23Execute these stages in order:
24
25### Stage 1: Voice Cleanup
26
27If the input appears to be voice transcription or casual speech, clean it up:
28
29**Remove:**
30- Filler words: "um", "uh", "like", "you know", "basically", "actually", "literally", "right?"
31- Self-corrections: "no wait", "I mean", "actually scratch that", "let me rephrase"
32- False starts: repeated or abandoned sentence beginnings
33- Verbal tics: "okay so", "let me think", "hmm"
34
35**Normalize:**
36- Convert spoken patterns to written form
37- Fix run-on sentences
38- Add proper punctuation
39
40### Stage 2: Intent Extraction
41
42Parse the cleaned input to identify:
43
44| Element | What to Find |
45|---------|--------------|
46| Core task | What does the user actually want done? |
47| Referenced files | Documents, project files, URLs mentioned |
48| Constraints | Limitations, requirements, boundaries |
49| Preferences | Style, format, approach preferences |
50| Success criteria | What does "done" look like? |
51| Scope | Single task vs. multi-step project |
52
53### Stage 3: Gap Detection
54
55Scan for missing or ambiguous elements:
56
57- [ ] Ambiguous requirements ("process them somehow")
58- [ ] Missing success criteria (no clear definition of "done")
59- [ ] Unclear scope (single task or multi-step?)
60- [ ] Format unspecified (code? prose? structured output?)
61- [ ] Model unknown (Opus 4.5 or Sonnet 4.5?)
62- [ ] Technical choices undefined (which library/approach/pattern?)
63
64**If gaps exist, ask ALL clarifying questions at once:**
65
66```
67Before I polish this prompt, I need to clarify a few things:
68
691. [First gap-specific question]
702. [Second gap-specific question]
713. Which Claude model are you using? (Opus 4.5 / Sonnet 4.5 / Not sure)
72```
73
74Wait for user response before proceeding.
75
76### Stage 4: Apply Best Practices
77
78Reference [reference.md](reference.md) for detailed guidelines. Apply these transformations:
79
80**Universal (Claude 4.x):**
81- Make instructions explicit and direct (no hints)
82- Add contextual framing (explain WHY, not just WHAT)
83- Use positive framing ("use prose" not "don't use bullets")
84- Structure with XML tags for complex inputs
85- Use action-oriented language ("Change this" not "Can you suggest")
86- Include chain of thought prompts for complex reasoning
87- Define clear success criteria
88- Use normal tone (avoid aggressive CAPS/MUST/CRITICAL)
89
90**If Opus 4.5:**
91- Add anti-over-engineering guardrail: "Keep solutions simple. Only make changes directly requested."
92- Avoid the word "think" → use "consider", "evaluate", "believe"
93- Leverage planning strength: "Build an editable plan before executing"
94- Keep prompts token-efficient (don't over-pad)
95- Provide policy/constraint context when relevant
96
97**If Sonnet 4.5:**
98- Break complex tasks into explicit steps
99- Optimize for speed/throughput
100- Add more explicit step-by-step guidance for intricate tasks
101
102### Stage 5: Structure the Prompt
103
104Generate the polished prompt using this template:
105
106```xml
107<context>
108[Background info, relevant project context, referenced files, WHY this task matters]
109</context>
110
111<task>
112[Clear, explicit statement of what needs to be done - action-oriented]
113</task>
114
115<requirements>
116[Specific constraints, boundaries, must-haves, technical choices]
117</requirements>
118
119<success_criteria>
120[What "done" looks like, how to verify completion]
121</success_criteria>
122
123<approach> <!-- Optional: include for complex tasks -->
124[Suggested approach OR "Build an editable plan before executing"]
125</approach>
126```
127
128**Formatting rules:**
129- Keep each section focused and concise
130- Use bullet points only when listing multiple items
131- Omit sections that aren't relevant to the task
132- For simple tasks, a streamlined version without XML tags is acceptable
133
134### Stage 6: Preview for Approval
135
136Present the polished prompt to the user:
137
138```
139Here's your polished prompt:
140
141---
142
143[POLISHED PROMPT CONTENT]
144
145---
146
147**Options:**
148- **Execute** - Run this prompt now
149- **Edit** - I'll modify something first
150- **Redo** - Start over with different clarifications
151```
152
153**Wait for user approval before executing.**
154
155### Stage 7: Execute or Iterate
156
157Based on user choice:
158- **Execute**: Immediately act on the polished prompt as if the user had typed it directly
159- **Edit**: User modifies the prompt, then re-preview
160- **Redo**: Return to Stage 1 with new input or Stage 3 for new clarifications
161
162## Important Notes
163
164- Always preview before executing - never skip the approval step
165- If the input is already well-structured, acknowledge this and offer minor polish only
166- Learn from user feedback - if they frequently edit a certain way, note the pattern
167- For very short/simple inputs, use a streamlined format without full XML structure
168- The goal is USEFUL prompts, not LONG prompts - be concise
169
170## See Also
171
172- [reference.md](reference.md) - Detailed Anthropic best practices for Claude 4.x, Opus 4.5, and Sonnet 4.5
173- [examples.md](examples.md) - Before/after transformation examples