sci-polish

Name: sci-polish
Author: ShZhao27208

167stars23forks

How to install sci-polish?

shzhao27208/aut_sci_write/sci-polish

$npx -y skills add ShZhao27208/Aut_Sci_Write --skill sci-polish

Installs into the current project.

Summary

Two-stage academic paper polishing skill. Stage A reduces AI detection traces (targeting GPTZero, Turnitin, Originality.ai). Stage B performs 8-dimension quality improvement (grammar, tone, coherence, conciseness, terminology, structure, argument clarity, journal compliance). Use

Files · 1

View on GitHub

SKILL.md

1	# Sci-Polish — Academic Paper Polishing & De-AI
2
3	Two-stage academic paper polishing: AI trace reduction + 8-dimension quality improvement.
4
5	---
6
7	## Hard Rules (NEVER violate)
8
9	1. NEVER modify: LaTeX commands, math environments (`$...$`, `\[...\]`, `equation`, `align`, etc.), `\cite{}`, `\ref{}`, `\label{}`, algorithm blocks, code listings, figure/table environments (only caption text is editable)
10	2. NEVER fabricate: data, citations, experimental results, baselines, metrics, author names
11	3. NEVER remove: author's technical claims without explicit permission
12	4. NEVER introduce: Tier 1 AI words (see `references/word-choice-anti-ai.md`) during Phase B edits
13	5. Mark as `[PENDING VERIFICATION]`: any claim you cannot verify from the text
14	6. Preserve meaning: the exact semantic content of every technical statement must be unchanged
15	7. Output format: revised text + polish report. No preamble or trailing changelog unless user asks
16	8. When uncertain about domain terminology: keep the original, flag for user review
17
18	---
19
20	## Execution Modes
21
22	\| Mode \| Command \| Behavior \|
23	\|------\|---------\|----------\|
24	\| `full` (default) \| "polish this paper" \| Phase A → Phase B (all 8 dimensions) \|
25	\| `deai-only` \| "just reduce AI rate" / "只降AI率" \| Phase A only \|
26	\| `polish-only` \| "just polish, skip de-AI" / "只润色" \| Phase B only (all 8 dimensions) \|
27	\| `polish-select` \| "fix grammar and coherence" \| Phase B, selected dimensions only \|
28
29	---
30
31	## Step 0: Input Processing
32
33	1. Identify input format:
34	- PDF → extract visible text (preserve section structure)
35	- LaTeX → extract visible paragraphs (protect all commands)
36	- Markdown / plain text → use directly
37	2. Identify protection zones: math environments, citations, labels, algorithm blocks, code blocks, figure/table environments
38	3. Ask target journal (optional): affects Dimension 8 scoring. If not specified, use general academic standards.
39	4. Confirm execution mode: full / deai-only / polish-only / polish-select
40
41	---
42
43	## Step 1: Phase A — De-AI Editing
44
45	Reference files:
46	- `references/ai-detection-patterns.md` — pattern library for scanning
47	- `references/de-ai-strategies.md` — rewriting strategies
48	- `references/word-choice-anti-ai.md` — word replacement table
49
50	### 1.1 Scan
51
52	Scan text against `ai-detection-patterns.md`. Categorize findings:
53	- High priority: Tier 1 words, template phrases, burstiness deficit (SD < 8), RLHF fingerprints
54	- Medium priority: Tier 2 threshold violations, structural symmetry, transition stacking
55	- Low priority: Tier 3 context-dependent, punctuation fingerprints, minor style uniformity
56
57	### 1.2 Rewrite (Priority Order)
58
59	Follow the strategy priority from `de-ai-strategies.md`:
60
61	```
62	a. Burstiness enforcement — restructure sentence lengths (Strategy 1)
63	Target: SD > 8, range ≥ 20, mid-band < 50%
64
65	b. Vocabulary de-AI — replace Tier 1 words, compress template phrases (Strategy 2)
66	Consult word-choice-anti-ai.md for replacements
67
68	c. Sentence restructuring — vary syntax patterns (Strategy 3)
69	No 3+ consecutive same-structure sentences
70
71	d. Concept concretization — add specifics where available (Strategy 4)
72	Replace vague claims with data from the paper itself
73
74	e. Argumentation enrichment — add depth to claims (Strategy 5)
75	Insert boundary conditions, alternative interpretations
76
77	f. Perplexity elevation — introduce unpredictability (Strategy 6)
78	Unconventional openers, domain-specific minor terms
79
80	g. Style disruption — break uniformity (Strategy 7)
81	Vary paragraph density and register between sections
82	```
83
84	### 1.3 Self-Audit
85
86	After rewriting, re-scan output against `ai-detection-patterns.md`:
87	- Any remaining High-priority patterns? → fix them
88	- Burstiness check: SD > 8? Range ≥ 20? Mid-band < 50%?
89	- Protection zones intact? (LaTeX commands, math, citations unchanged)
90	- Meaning preserved? (no semantic drift from original)
91
92	---
93
94	## Step 2: Phase B — 8-Dimension Polish
95
96	Reference file: `references/polish-dimensions.md`
97
98	### 2.1 Score All Dimensions (1-5)
99
100	Rate each dimension based on the checkpoints in `polish-dimensions.md`:
101	1. Grammar & Syntax
102	2. Academic Tone
103	3. Coherence & Flow
104	4. Conciseness
105	5. Terminology
106	6. Structure & Formatting
107	7. Cla

People who liked this also viewed

6 picks

Type

sci-polish

167stars23forks

View on GitHub

How to install sci-polish?

shzhao27208/aut_sci_write/sci-polish

$npx -y skills add ShZhao27208/Aut_Sci_Write --skill sci-polish

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Sci-Polish — Academic Paper Polishing & De-AI
2
3	Two-stage academic paper polishing: AI trace reduction + 8-dimension quality improvement.
4
5	---
6
7	## Hard Rules (NEVER violate)
8
9	1. NEVER modify: LaTeX commands, math environments (`$...$`, `\[...\]`, `equation`, `align`, etc.), `\cite{}`, `\ref{}`, `\label{}`, algorithm blocks, code listings, figure/table environments (only caption text is editable)
10	2. NEVER fabricate: data, citations, experimental results, baselines, metrics, author names
11	3. NEVER remove: author's technical claims without explicit permission
12	4. NEVER introduce: Tier 1 AI words (see `references/word-choice-anti-ai.md`) during Phase B edits
13	5. Mark as `[PENDING VERIFICATION]`: any claim you cannot verify from the text
14	6. Preserve meaning: the exact semantic content of every technical statement must be unchanged
15	7. Output format: revised text + polish report. No preamble or trailing changelog unless user asks
16	8. When uncertain about domain terminology: keep the original, flag for user review
17
18	---
19
20	## Execution Modes
21
22	\| Mode \| Command \| Behavior \|
23	\|------\|---------\|----------\|
24	\| `full` (default) \| "polish this paper" \| Phase A → Phase B (all 8 dimensions) \|
25	\| `deai-only` \| "just reduce AI rate" / "只降AI率" \| Phase A only \|
26	\| `polish-only` \| "just polish, skip de-AI" / "只润色" \| Phase B only (all 8 dimensions) \|
27	\| `polish-select` \| "fix grammar and coherence" \| Phase B, selected dimensions only \|
28
29	---
30
31	## Step 0: Input Processing
32
33	1. Identify input format:
34	- PDF → extract visible text (preserve section structure)
35	- LaTeX → extract visible paragraphs (protect all commands)
36	- Markdown / plain text → use directly
37	2. Identify protection zones: math environments, citations, labels, algorithm blocks, code blocks, figure/table environments
38	3. Ask target journal (optional): affects Dimension 8 scoring. If not specified, use general academic standards.
39	4. Confirm execution mode: full / deai-only / polish-only / polish-select
40
41	---
42
43	## Step 1: Phase A — De-AI Editing
44
45	Reference files:
46	- `references/ai-detection-patterns.md` — pattern library for scanning
47	- `references/de-ai-strategies.md` — rewriting strategies
48	- `references/word-choice-anti-ai.md` — word replacement table
49
50	### 1.1 Scan
51
52	Scan text against `ai-detection-patterns.md`. Categorize findings:
53	- High priority: Tier 1 words, template phrases, burstiness deficit (SD < 8), RLHF fingerprints
54	- Medium priority: Tier 2 threshold violations, structural symmetry, transition stacking
55	- Low priority: Tier 3 context-dependent, punctuation fingerprints, minor style uniformity
56
57	### 1.2 Rewrite (Priority Order)
58
59	Follow the strategy priority from `de-ai-strategies.md`:
60
61	```
62	a. Burstiness enforcement — restructure sentence lengths (Strategy 1)
63	Target: SD > 8, range ≥ 20, mid-band < 50%
64
65	b. Vocabulary de-AI — replace Tier 1 words, compress template phrases (Strategy 2)
66	Consult word-choice-anti-ai.md for replacements
67
68	c. Sentence restructuring — vary syntax patterns (Strategy 3)
69	No 3+ consecutive same-structure sentences
70
71	d. Concept concretization — add specifics where available (Strategy 4)
72	Replace vague claims with data from the paper itself
73
74	e. Argumentation enrichment — add depth to claims (Strategy 5)
75	Insert boundary conditions, alternative interpretations
76
77	f. Perplexity elevation — introduce unpredictability (Strategy 6)
78	Unconventional openers, domain-specific minor terms
79
80	g. Style disruption — break uniformity (Strategy 7)
81	Vary paragraph density and register between sections
82	```
83
84	### 1.3 Self-Audit
85
86	After rewriting, re-scan output against `ai-detection-patterns.md`:
87	- Any remaining High-priority patterns? → fix them
88	- Burstiness check: SD > 8? Range ≥ 20? Mid-band < 50%?
89	- Protection zones intact? (LaTeX commands, math, citations unchanged)
90	- Meaning preserved? (no semantic drift from original)
91
92	---
93
94	## Step 2: Phase B — 8-Dimension Polish
95
96	Reference file: `references/polish-dimensions.md`
97
98	### 2.1 Score All Dimensions (1-5)
99
100	Rate each dimension based on the checkpoints in `polish-dimensions.md`:
101	1. Grammar & Syntax
102	2. Academic Tone
103	3. Coherence & Flow
104	4. Conciseness
105	5. Terminology
106	6. Structure & Formatting
107	7. Cla

People who liked this also viewed

6 picks

Type