cover-letter

Submission cover-letter assistant for existing LaTeX manuscripts. It generates, optimizes, and checks editor-facing letters while keeping every strong letter claim tied to manuscript evidence.

Use It For

• Generate a cover letter draft from main.tex.
• Optimize an existing cover_letter.md or cover_letter.tex without overwriting it.
• Align-check novelty, contribution, numeric, and scope claims against the manuscript.
• Check whether the letter is framed for a bundled journal or conference template.
• Run final declaration, length, cliché, AI-tone, and paragraph-shape checks.

Do Not Use It For

• Editing the manuscript source; use latex-paper-en or latex-thesis-zh.
• Full reviewer-style manuscript critique; use paper-audit.
• Searching bibliography files; use bib-search-citation.
• Typst manuscripts; this version supports LaTeX manuscripts only.
• Rebuttal or response-to-reviewer letters.

Mode Router

Mode	Use when	Primary command
generate	You have a manuscript but no letter draft	uv run python academic-writing-skills/cover-letter/scripts/cover_letter.py --mode generate --manuscript main.tex --journal nature --json
optimize	You have a draft and want safer framing	uv run python academic-writing-skills/cover-letter/scripts/cover_letter.py --mode optimize --letter cover_letter.md --manuscript main.tex --journal nature --json
align-check	You need to verify claims against the paper	uv run python academic-writing-skills/cover-letter/scripts/cover_letter.py --mode align-check --letter cover_letter.md --manuscript main.tex --json
journal-fit	You need venue fit scoring	uv run python academic-writing-skills/cover-letter/scripts/cover_letter.py --mode journal-fit --letter cover_letter.md --journal nature --json
presubmission	You need final mechanical checks	uv run python academic-writing-skills/cover-letter/scripts/cover_letter.py --mode presubmission --letter cover_letter.md --journal nature --json

Supported bundled venues: nature, science, cell, ieee-trans, acm, springer-lncs, neurips, icml, cvpr, and generic.

Minimum Inputs

• --manuscript main.tex for generate and align-check.
• --letter cover_letter.md or --letter cover_letter.tex for optimize, align-check, journal-fit, and presubmission.
• --journal <venue> for journal-fit and venue-specific pre-submission rules.
• --json when you want a machine-readable issue bundle.

Script Entry Points

Script	Purpose
cover_letter.py	Unified public CLI for all five modes
extract_manuscript_facts.py	Deterministic facts extraction from the manuscript
build_letter_claim_map.py	Cover-letter claim inventory
align_check.py / verify_letter_against_manuscript.py	Claim-evidence verification
journal_fit_check.py	Venue-fit scoring
presubmission_check.py	Declaration, length, cliché, and tone checks

Output Artifacts

• generate returns manuscript facts plus a deterministic draft scaffold with placeholders for unknown editor or declaration fields.
• optimize returns line-anchored suggestions and claim-evidence warnings; it does not overwrite the draft.
• align-check returns unsupported, over-scoped, or missing-evidence claim findings.
• journal-fit returns HIGH / MEDIUM / LOW axis verdicts and mapped findings.
• presubmission returns mechanical findings for required declarations, length, cliché, tone, and paragraph shape.

Findings use severity, priority, source_kind, and comment_type; script-backed findings are intended to be rerunnable.

Public Resources

References

• AI disclosure policy
• Claim-evidence contract
• Forbidden phrases
• Issue schema
• Journal tiers
• Letter structure
• Mode guide
• Pre-submission rules

Templates

• ACM
• Cell
• CVPR
• Generic
• ICML
• IEEE Transactions
• Nature
• NeurIPS
• Science
• Springer LNCS

Examples

• Align check only
• Generate a Nature letter
• CVPR vs. TPAMI journal fit
• Optimize and align

Agent Contracts

• Claims-evidence reviewer
• Committee editor

Common Requests

Generate a Nature cover letter from main.tex and include placeholders where metadata is missing.

Optimize cover_letter.md for an IEEE Transactions submission, but do not edit the file directly.

Check whether every strong claim in this cover letter is supported by main.tex.

Score this letter for NeurIPS journal-fit style framing and tell me the weakest axis.

Run final pre-submission checks on cover_letter.md before I paste it into the submission system.