sec-conf-paper

name

sec-conf-paper

description

Use when generating a security conference paper draft that should preserve the existing sec-conf-write and sec-conf-plot workflows, produce a full English draft with figures first, then produce a Chinese text-only mirror draft while leaving figure assets unchanged. Use for bilingual ACM CCS, NDSS, IEEE S&P, USENIX Security, RAID, ACSAC, and similar security-paper drafting workflows where the user provides fixed English and Chinese headings.

Security Conference Paper

Purpose

Use this skill to orchestrate a bilingual security-conference paper workflow without replacing the existing specialized skills.

The intended workflow is:

1. Use sec-conf-write to generate or revise the full English paper draft.
2. Use sec-conf-plot to generate publication-style English figures, diagrams, and plots.
3. Create a Chinese mirror draft by translating the paper text only.
4. Keep all figure assets, generated plotting scripts, image labels, source data, and figure files unchanged.

This skill is for paper drafts, not standalone CVE advisories, pentest reports, exploit notes, or vulnerability reports.

Required Skill Coordination

When this skill triggers for a full paper draft, also use:

• sec-conf-write for the English manuscript narrative, section drafting, reviewer-facing checks, threat model, evaluation framing, limitations, ethics, and related work.
• sec-conf-plot for figures, charts, diagrams, visual QA, and reproducible figure artifacts.

Do not weaken or rewrite the rules in those skills. Treat this skill as the outer workflow that decides ordering and bilingual handling.

Title Lock

The user owns all headings and titles.

• Preserve user-provided English headings exactly in the English draft.
• Preserve user-provided Chinese headings exactly in the Chinese mirror draft.
• Do not translate, rewrite, rename, reorder, or invent section titles when the user supplies a title plan.
• If a required title is missing and the user has said they will provide titles, leave a clear placeholder such as [TITLE PROVIDED BY USER] rather than inventing one.
• Body text may be drafted, rewritten, and translated; headings are locked.

Workflow

1. Gather Inputs

Identify the available materials:

• Fixed English heading plan.
• Fixed Chinese heading plan, if already provided.
• Existing manuscript files, notes, experiments, tables, logs, datasets, diagrams, and related-work notes.
• Target venue, if known.
• Expected output format, usually LaTeX or Markdown.

Ask only for missing information that blocks factual correctness: attacker model, protected asset, vulnerability condition, dataset, evaluation result, disclosure status, or fixed headings.

2. Generate English Draft

Use sec-conf-write normally.

Requirements:

• Produce the full English manuscript first.
• Preserve locked English headings.
• Keep claims grounded in supplied evidence.
• Mark unknown citations, measurements, implementation facts, or disclosure outcomes as [CITATION NEEDED], [RESULT NEEDED], [DETAIL NEEDED], or [DISCLOSURE STATUS NEEDED].
• Do not fabricate CVEs, datasets, baselines, measurements, venue rules, citations, or artifact claims.

3. Generate Figures

Use sec-conf-plot normally after the English draft has enough stable RQ, method, and evaluation content.

Requirements:

• Keep figure labels and visual language in English unless the user explicitly asks otherwise.
• Generate reproducible figure sources for data plots, such as figures/gen_fig_<name>.py.
• Export vector PDF plus PNG previews when appropriate.
• Preserve raw data or point to the source table/log in figure generation scripts.
• Do not create decorative security imagery.

4. Create Chinese Text Mirror

Create a separate Chinese draft after the English draft and figures exist.

Translate all manuscript text that is outside immutable technical syntax:

• Abstract/body paragraphs.
• Captions and table notes when they are manuscript text.
• Footnotes, appendix prose, ethics prose, limitations, and acknowledgments.
• Reviewer-facing comments or TODO explanations, if present.

Keep these unchanged:

• Figure files and all content embedded inside images/PDF figures.
• Plotting scripts and source data.
• LaTeX commands, environments, package commands, and macro names.
• Citation keys such as \cite{key} and bibliography entries unless explicitly asked.
• Labels, refs, anchors, file paths, URLs, code identifiers, API names, CVE IDs, project names, dataset IDs, equations, math notation, and algorithm symbols.
• User-provided Chinese headings.

The Chinese draft is a mirror for human refinement and logic review. It must not overwrite the English manuscript.

5. Synchronization Checks

Before finishing, check:

• English draft exists and uses the locked English headings.
• Chinese mirror draft exists and uses the locked Chinese headings.
• Section order and paragraph-level content correspond between English and Chinese drafts.
• Figure references in the Chinese draft point to the same figure assets as the English draft.
• Figure files and plot-generation scripts were not modified during translation.
• All unresolved factual gaps remain explicitly marked.

If the workspace is a git repository, use git diff -- figures or equivalent inspection to confirm translation did not alter figure assets.

Output Convention

Prefer preserving the project format:

• If the paper is LaTeX, create or update English .tex files and a separate Chinese mirror file or folder, such as paper_zh.tex or zh/.
• If the paper is Markdown, create a separate Chinese mirror file, such as paper_zh.md.
• Keep figure outputs under the existing figures/ convention.

Final responses should report:

• English draft path.
• Chinese mirror draft path.
• Figure directory and major generated figure files.
• Any unresolved [CITATION NEEDED], [RESULT NEEDED], or [DETAIL NEEDED] items.