polish
Perform final pre-submission checks: formatting, citations, figures, metadata, and build verification. Systematic checklist identifies remaining issues and suggests fixes.
- License 许可
- MIT
- 最近更新
- 2026-10-23
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- polish
- description
- Perform final pre-submission checks: formatting, citations, figures, metadata, and build verification. Systematic checklist identifies remaining issues and suggests fixes.
03 SKILL.md 正文
Submission Polish
Perform a final quality check on a research paper before submission. This is the last line of defense: catch everything that would annoy a reviewer or cause a desk rejection.
Step 1: Read Context
Read .papermill/state.md (Read tool) for:
- Venue: Target venue and its requirements.
- Format: Paper format (latex, markdown, rmarkdown).
- Review history: Outstanding issues from previous reviews.
If .papermill/state.md does not exist, the pre-flight check can still run: infer the format from the manuscript files and ask the user for the target venue. The checklist works regardless. Suggest running /papermill:init to capture venue and format persistently.
Read the complete manuscript (Read tool).
Step 2: Pre-Flight Checklist
Work through each category systematically. Report issues as they are found.
Formatting
- [ ] Paper compiles without errors
- [ ] No LaTeX warnings (especially undefined references, missing citations)
- [ ] Page count within venue limits
- [ ] Correct venue template/style file used
- [ ] Margins, font size, and spacing match requirements
- [ ] Line numbers included if required by venue
- [ ] Anonymous submission if required (no author names in text or metadata)
Abstract
- [ ] Word count within venue limits (typically 150-300 words)
- [ ] Self-contained (no citations, no undefined terms)
- [ ] States the problem, approach, main result, and significance
- [ ] Matches the paper's actual content
Citations and References
- [ ] All citations resolve (no "?" in the PDF)
- [ ] Bibliography is complete (no missing fields)
- [ ] Citation style matches venue requirements
- [ ] All references are cited in the text (no orphan bib entries that should be removed)
- [ ] All cited works are in the bibliography
- [ ] Self-citations are reasonable in number
Figures and Tables
- [ ] All figures are referenced in the text
- [ ] Figure captions are self-contained (reader should understand without reading the text)
- [ ] Resolution is sufficient for print (300+ DPI for raster images)
- [ ] Color is used accessibly (works in grayscale, colorblind-safe)
- [ ] Tables have clear headers and units
Mathematical Content
- [ ] All symbols are defined before use
- [ ] Notation is consistent throughout
- [ ] Equation numbering is correct and referenced
- [ ] Theorem/definition/proof environments are properly labeled
- [ ] No orphan proofs (every proof has a parent theorem/proposition)
Writing Quality
- [ ] No first-person narrative inconsistencies ("I" vs "we")
- [ ] Consistent tense usage
- [ ] No obvious typos or grammatical errors
- [ ] Acronyms defined on first use
- [ ] No overly long paragraphs (break up walls of text)
Metadata
- [ ] Author names and affiliations are correct
- [ ] ORCID identifiers included if venue supports them
- [ ] Keywords/subject classification provided
- [ ] Acknowledgments section present (funding, help, etc.)
- [ ] Data availability statement if required
- [ ] Code availability statement if applicable
Supplementary Material
- [ ] Appendices are referenced from the main text
- [ ] Supplementary files are in the required format
- [ ] Appendix content is not essential to understanding the main paper
Step 3: Build Verification
Launch the format validator (Agent tool, subagent_type: "papermill:format-validator") with the manuscript path from Step 1 in <paper> and the state file contents in <state>. It owns the build recipe for every format and reports build status, page count, warnings, undefined references, and overfull/underfull boxes. Fold its findings into the checklist results.
Step 4: Present Results
Present the checklist results:
Pre-Flight Report
Status: [Ready / Issues found]
Passing: N/M checks passed Issues:
- [issue + suggested fix]
- [issue + suggested fix] ...
Step 5: Fix Issues
Offer to fix issues directly:
- Formatting fixes: Apply them.
- Content fixes: Suggest specific changes and ask for approval.
- Style fixes: Propose rewording.
Step 6: Update State File
After all issues are resolved (if .papermill/state.md does not exist, skip this step and suggest /papermill:init):
- Set
stagetosubmissionin.papermill/state.md(Edit tool). - Append a timestamped note: "Pre-flight check passed. Ready for submission to [venue]."
Step 7: Suggest Next Steps
Paper is polished and ready. Final steps:
- Submit: Follow the venue's submission instructions.
/papermill:venue: If you haven't selected a venue yet, do so now.- Archive: Consider creating a git tag marking the submitted version.
04 同仓库更多 Skills
worldsmith-methodology
Use when working in a project with a worldbuilding documentation ecosystem (docs/ or lore/ directory). Provides the documentation-first methodology: canonical hierarchy, doc roles (timeline, lore, characters, systems), propagation discipline, and the dual canonical and exploratory workflow.
prose-craft
Use when writing fiction, editing fiction prose, reviewing chapters, or drafting scenes. Covers prose craft, dialogue mechanics, scene structure, and common failure modes in AI-assisted fiction. Trigger phrases: "write fiction", "edit chapter", "prose review", "scene draft", "dialogue check", "fiction craft", "edit this scene", "review my prose".
vista
This skill should be used when the user asks about research directions, open problems, future work, or follow-up research from the academic literature. Trigger phrases include "open problems in X", "what's next for Y", "future work for paper Z", "research directions on T", "salient follow-up research", "what should I work on next in", "find me follow-up research unrelated to my prior work", "broad survey of W", "neglected directions in V". Routes the request to the right Vista MCP tool, reads back the structured paper sections, and synthesizes research directions in the conversation.
synthesize
Make a research-agent run conclude and write synthesis.md (outcome, key findings, failed approaches, open questions, recommendations). Signals a live researcher to wrap up, or launches one in synthesis mode when none is running. Use when results are good enough or the agent is stalling.
status
Show the current state of a research-agent run from .research/: liveness, cycles and budget, sub-problem and hypothesis statuses, eval trend, current focus, and branches. Read-only.
resume
Resume an interrupted research-agent run from .research/. Re-launches the researcher to reload its structured state and continue from the current focus, optionally on a named branch or with a new budget. Use after a session restart, crash, or explicit pause.