runtsang/RebuttalStudio / skills/stage2/template
stage2-refine-template
VerifiedBase template for Stage2 rebuttal refinement. Use when creating or extending conference-specific Stage2 skills so all venues share the same high-quality writing constraints, JSON output format, and normalization rules.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.8 KB |
SKILL.md
View on GitHub---
name: stage2-refine-template
description: Base template for Stage2 rebuttal refinement. Use when creating or extending conference-specific Stage2 skills so all venues share the same high-quality writing constraints, JSON output format, and normalization rules.
---
# Stage 2 · Refine Base Template
Use this template as the shared core for all Stage2 conference variants.
Conference-specific files must apply this template first, then add only necessary overrides.
## Goal
Turn a rough outline into reviewer-facing rebuttal prose that:
1. Directly addresses the reviewer concern.
2. Keeps claims factual and bounded by provided evidence.
3. Uses concise, polite, academically confident style.
4. Optionally starts with one courteous opener.
## Input
- `response_id` / `title` / `source` / `source_id` (metadata)
- `quoted_issue` (verbatim reviewer concern)
- `draft` or `outline` (user-provided content)
- optional style hints (tone, length, verbosity)
## Output
Return strict JSON only:
```json
{ "draft": "> **Reviewer's Comment**: [quoted_issue]\n\n**Response**: [polished rebuttal prose]" }
```
## Core writing rules
1. **Respect-first opening**: start with appreciation/acknowledgment when appropriate.
2. **Issue anchoring**: connect to `quoted_issue` within the first 1–2 sentences.
3. **Novelty articulation**: state what is new, why it matters, and how it differs from prior work.
4. **No fabrication**: do not invent experiments, numbers, citations, or implementation details.
5. **Constructive tone**: avoid defensive language.
6. **Concise precision**: keep paragraphs short and information-dense.
7. **Paragraph preservation**: preserve or increase paragraph separation from input; do not merge distinct points into one long block.
## Courtesy opener bank
Select exactly one opener when an opener is needed:
- "Thank you for the excellent suggestion."
- "Thank you for pointing this out."
- "Thank you for highlighting this important point about [TOPIC]."
- "We appreciate the opportunity to further articulate the novelty of [METHOD/IDEA]."
- "Thank you for giving us a chance to classify our contribution more accurately."
### Opener selection guidance
- Actionable improvement advice -> "Thank you for the excellent suggestion."
- Omission or ambiguity pointed out -> "Thank you for pointing this out."
- Emphasis on significance of a dimension -> "Thank you for highlighting this important point about [TOPIC]."
- Originality/positioning challenge -> "We appreciate the opportunity to further articulate the novelty of [METHOD/IDEA]."
## Markdown normalization rules
If input contains structured content, normalize in final `draft`:
1. **Tables/charts -> Markdown table**
2. **Code -> fenced code block** (with language tag if known)
3. **Formula -> Markdown math**
- Inline math: `$...$`
- Display math: `$$...$$` on separate lines
## Shared refine workflow
1. Read `quoted_issue` and identify reviewer intent.
2. Extract factual points from `draft`/`outline`; remove unsupported claims.
3. Choose one opener from the opener bank if needed.
4. Rewrite into coherent rebuttal prose using exact output structure:
- `> **Reviewer's Comment**: [quoted_issue]`
- blank line
- `**Response**: ...`
5. Normalize table/code/formula content.
6. Return strict JSON only.
## Hard constraints
- Do not alter the meaning of user-provided technical claims.
- Do not introduce new citations unless already provided.
- Do not over-promise without evidence.
- Keep output ready for direct rebuttal editing.
- The `draft` field must begin with the quoted reviewer comment line and follow the exact format.
## Conference extension contract
Conference-specific Stage2 files may override only what is truly different:
- conference naming/labeling
- extra mandatory phrase variants (if a venue requires them)
- venue-specific style constraints
Keep all other rules inherited from this template.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |