Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mantou6666/Math-Modeling-Agent-Flow --skill math-modeling-finalizergit clone --depth 1 https://github.com/mantou6666/Math-Modeling-Agent-FlowWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mantou6666/math-modeling-agent-flow/math-modeling-finalizer)<a href="https://agentmods.dev/skills/mantou6666/math-modeling-agent-flow/math-modeling-finalizer"><img src="https://agentmods.dev/badge/skills/mantou6666/math-modeling-agent-flow/math-modeling-finalizer.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00160 | $0.03211 |
| Opus 5 | $0.00080 | $0.01605 |
| Sonnet 5 | $0.00032 | $0.00642 |
| Haiku 4.5 | $0.00016 | $0.00321 |
Grade A, and why
math-modeling-finalizer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 158 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Math Modeling Finalizer v2.1.0-rc2
本 Skill 只负责已经基本完成的项目怎样可靠收口。它不重新拥有数学事实,不建立第二套 Result/Claim Registry,也不通过固定 A—H 状态机强迫所有项目走相同流程。
默认围绕四项用户价值:
- 参赛者尺度、行为保持的代码人味化;
- 最终完整复审与最小 route-back;
- 从真实 AI ledger 生成 AI 工具使用详情 DOCX;
- 支撑材料 ZIP、DOCX/PDF 与最终交付检查。
0. 输入与事实边界
优先读取:PROJECT_CONTEXT.md、solver_handoff.json、Result Certificates、paper_handoff.json、formula_registry.json、references.bib/citation audit、ai_use_ledger.jsonl、当前论文/DOCX/PDF 和 Contest/Team Profile。 对 paper_handoff.json 不做盲目信任:进入 Freeze 前先运行 python scripts/verify_upstream_paper_handoff.py <paper_handoff.json>;任何越界路径、缺失文件或 hash 不匹配均阻断受影响的最终化 gate。
科学事实只来自 Solver Result Certificate/canonical formal output。Paper Handoff 负责当前稿件、Claim/引用/公式映射。Finalization 只验证这些链是否一致。
数学错误 → route_back.json owner=SOLVER_*;写作/引用/公式语义错误 → owner=PAPER_*;纯排版/交付问题留在 Finalization。
0.1 Freeze → Audit → Repair-safe → Revalidate → Submit
把收口视为五个可回查的决策门,而不是盲目执行的固定流水线:
- Freeze:记录 Solver/Paper Handoff、Result Certificate、代码/文稿/PDF/ledger 的 path、SHA-256、profile 和运行配置。任何 hash 变化都使对应下游通过证据失效。
- Audit:只检查事实链、行为链、排版链和交付链;区分
PASS、REVIEW、FAIL和NOT_VERIFIED。没有 exact PDF 时保留未验证状态。 - Repair-safe:根据 owner 选择最小修改范围。Finalization 不改 Problem Contract、formal evaluator、正式结果、模型结论或 Paper 的科学 claim;发现这些问题就 route-back。
- Revalidate:修复后只重跑受影响的最小依赖链,再跑最终交付 gate。代码 hash、Result Certificate、Evidence Registry、Paper Handoff 和 ZIP manifest 必须重新对齐。
- Submit:只有阻塞项清零、未验证项显式披露、exact PDF 已视觉检查(若赛事要求)且人工选择/批准状态满足 Profile 时,才可标记交付通过。Finalization 不代表已替用户向平台提交。
每一门都要留下报告或结构化状态;不存在证据时写 NOT_VERIFIED,不得用“整体看起来没问题”替代 gate。
0.2 Portable Project Contract
Finalizer 继承 Solver 的 PROJECT_ROOT 相对路径契约。所有会进入 review/report/manifest/ZIP 的引用只保存项目相对路径和 hash;绝对路径只允许在运行时用于打开文件。正式输入若仍位于项目外,应先导入 inputs/、paper/ 或 final/ 再冻结。这样同一项目目录可以整体复制或压缩,换机器解压后无需重写 artifact 引用。
代码人味化仍是本 Skill 的默认亮点,不因 Public Edition 而移除;但任何改写都继续受行为回归与 Result Certificate 约束。
0.3 Runtime capability、降级与恢复
收口阶段尤其容易依赖 PDF 视觉、Office 渲染、代码执行与文件系统,因此先根据宿主实际能力判断 AVAILABLE / UNAVAILABLE / UNKNOWN,不写死某个模型/工具,也不通过对正式文件反复触发 unsupported 错误来探测。
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- AGENTS.md 387 B
- agents/openai.yaml 796 B
- CHANGELOG.md 2.1 KB
- CLAUDE.md 399 B
- evals/negative-cases.json 1.6 KB
- evals/trigger_cases.csv 1.3 KB
- INSTALL.md 470 B
- KNOWN_LIMITATIONS.md 649 B
- LICENSE 1.1 KB
- PACKAGE_MANIFEST.json 13 KB
- profiles/generic-contest.json 1.3 KB
- prompts/first-run-codex.md 680 B
- README.md 1.7 KB
- references/ai-disclosure-team-policy.md 1.2 KB
- references/ai-usage-details-docx.md 545 B
- references/capability-boundaries.md 2.2 KB
- references/competition-paper-visual-gate.md 1.4 KB
- references/evidence-levels.md 858 B
- references/generic-audit-checklist.md 731 B
- references/integration-contract.md 1.1 KB
- references/severity-and-actions.md 540 B
- references/student-scale-code-simplification.md 2.3 KB
- RELEASE_CHECKLIST.md 1.5 KB
- requirements.txt 54 B
- schemas/behavior-contract.schema.json 1.8 KB
- schemas/code-simplification.schema.json 1.2 KB
- schemas/delivery-checklist.schema.json 1.5 KB
- schemas/evidence-ref.schema.json 2.1 KB
- schemas/inventory.schema.json 1.5 KB
- schemas/options.schema.json 244 B
- schemas/README.md 256 B
- schemas/route-back.schema.json 1.0 KB
- schemas/section-budget.schema.json 2.3 KB
- schemas/visual-review.schema.json 539 B
- scripts/_common.py 7.0 KB runs code
- scripts/_safe_docx.py 4.3 KB runs code
- scripts/_trust.py 672 B runs code
- scripts/archive_audit.py 533 B runs code
- scripts/audit_docx_privacy.py 3.6 KB runs code
- scripts/audit_formula_fragments.py 3.2 KB runs code
- scripts/audit_page_number_fields.py 3.6 KB runs code
- scripts/audit_pdf_security.py 3.7 KB runs code
- scripts/audit_table_geometry.py 4.6 KB runs code
- scripts/build_growth_handoff.py 3.3 KB runs code
- scripts/build_package_manifest.py 1.4 KB runs code
- scripts/build_review_packet.py 2.1 KB runs code
- scripts/build_support_archive.py 4.9 KB runs code
- scripts/check_archive_report.py 2.4 KB runs code
- scripts/compare_csv.py 2.4 KB runs code
- scripts/compare_json_semantic.py 3.4 KB runs code
- scripts/generate_ai_usage_details_docx.py 2.8 KB runs code
- scripts/lint_audit_language.py 1.5 KB runs code
- scripts/paper_report.py 5.5 KB runs code
- scripts/quick_validate.py 1.5 KB runs code
- scripts/render_with_fallback.py 8.9 KB runs code
- scripts/self_test.py 9.0 KB runs code
- scripts/validate_ai_ledger.py 1.6 KB runs code
- scripts/validate_evals.py 766 B runs code
- scripts/validate_visual_review.py 3.2 KB runs code
- scripts/verify_archive.py 9.8 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 158 lines · 160 tokens per session scan A 45c4f58a2b92
math-modeling-finalizer is a skill published in the GitHub repository mantou6666/Math-Modeling-Agent-Flow (16 stars, last pushed 18d ago), licensed MIT. It adds 160 tokens to every session and 3,211 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
extracting-lab-tables
Detects and extracts tabular laboratory panels from PDFs, scans, and images into structured rows ready for OpenMed and FHIR. Use when the user has a CBC, CMP, lipid panel, or other lab report as a scanned image / PDF / spreadsheet and needs the test name, value, unit, reference range, and abnormal flag as clean rows.…
paper-length-gate
Deterministic artifact-backed manuscript readiness gate for meta-paper-write. Validates the workspace LaTeX artifact before compilation while leaving final page-count enforcement to compilepdf.
paper-plot-stub
Plot a results CSV (x, ybaseline, yours) as a two-line matplotlib chart and write a PDF. Demo-only.
latex-compile
Compile a LaTeX document and fix every error plus aesthetic issue (overfull/underfull boxes, widows, alignment, fonts) for a clean PDF and log. Use this instead of running pdflatex/latexmk manually — it avoids the latexmk stale-log trap and silent grep failures on binary log output, and it reformats rather than…
figure-drawing
MANDATORY skill for creating any figure, diagram, chart, table visualization, or TikZ graphic in a paper. Must be activated BEFORE writing any drawing code. Produces standalone .tex → compiled PDF/PNG → \includegraphics in paper. Never inline TikZ code directly into paper sections.
PDF files: create, read, merge, fill, OCR, edit text.