agentsclimarketplace

Thesis citation extract

Skill VictorAurelius/claude-starter-kit/skills/quality/thesis-citation-extract

Dùng khi user nói 'extract citations', 'trích dẫn luận văn', 'audit thesis cite', 'verify bibliography', 'orphan citation check', 'kiểm tra cite luận văn', hoặc khi review thesis chapter trước defense. Parse `[N]` cite keys từ thesis chapters + verify chéo với `documents/<thesis-dir>/references/bibliography.md` → báo 3 bucket: matched / orphan-body (cite không có entry) / orphan-bib (entry không cite).From its SKILL.md

Install
npx -y skills add VictorAurelius/claude-starter-kit --skill thesis-citation-extract

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
  • runs commandsInstructs the agent to run 3 commands, including `bash scripts/extract-citations.sh <chapter-files...>` and 2 more.

SKILL.md

4.8 KB, ~1.3k tokens by cl100k_base, as published. Nobody here has run it

Thesis Citation Extract Skill

Tự động extract [N] citation keys từ thesis chapter markdown + verify chéo với canonical bibliography (documents/<thesis-dir>/references/bibliography.md) theo IEEE format. Phát hiện 2 loại orphan: broken refs (cite trong body nhưng không có entry) + dead weight (entry trong bib nhưng không được cite). Output JSON cache vào data/last-run.json để track delta giữa các lần audit.

Khi nào dùng

  • Trước khi flip documents/<thesis-dir>/** chapter PR sang DONE
  • Audit thesis V1/V2/V3 trước defense GVHD/GVPB
  • Thesis closure sub-task — verify cite hygiene
  • Phát hiện citation drift sau khi user thêm/sửa bibliography

Quy trình (≈3 phút)

  1. Extract — chạy bash scripts/extract-citations.sh <chapter-files...> → output filename:line:cite_key cho mọi [N] / [N, M] / [N]–[M] trong body. Range expand thành cá nhân ([5]–[7] → 5, 6, 7); danh sách [1, 4, 9] tách 1+4+9.
  2. Verify — chạy bash scripts/verify-citations.sh <chapter-files...> <bibliography-path> → so sánh tập cite (body) vs tập entry (bib). Output 3 nhóm + báo cáo summary stdout + ghi JSON cache vào data/last-run.json.
  3. Interpret — đọc summary:
    • matched: N — cite có entry, OK
    • ⚠️ orphan-body: N — cite trong body nhưng KHÔNG có entry bib → broken reference (fix bib hoặc xóa cite)
    • ⚠️ orphan-bib: N — entry trong bib nhưng KHÔNG được cite → dead weight (xóa entry hoặc thêm cite trong body)
  4. Self-testbash scripts/self-test.sh validate logic trên 3 fixture (good / bad-orphan-body / bad-orphan-bib).

Skill contents

  • reference/ieee-citation-rules.md — quy ước IEEE in-text + bibliography flat list format (project canonical)
  • reference/orphan-detection.md — định nghĩa 2 loại orphan + decision tree fix
  • scripts/extract-citations.sh — parser regex \[[0-9]+\] với range expansion
  • scripts/verify-citations.sh — set-diff body cites vs bibliography entries
  • scripts/self-test.sh — synthetic fixtures verify extract + verify logic
  • data/last-run.json — cache lần verify gần nhất (delta tracking, generated at runtime)

Gotchas

  • Bibliography flat list — recommended convention: bibliography renumbered sang single sequential [1]...[N] flat (NOT per-chapter sections). Section headings ## Chapter N chỉ là grouping cho reader; cite key độc lập section.
  • Range expansion[5]–[7] dùng em-dash (U+2013), KHÔNG phải hyphen -. Parser support cả 2 (em-dash + hyphen) để phòng typo.
  • List separator[1, 2] dùng comma + space; parser tolerate [1,2] (no space) và [1 ,2] (trailing space).
  • Code block exclusion — KHÔNG parse [N] trong code fence ``` (vd array[1] Java code). Parser skip lines giữa code fence markers.
  • Markdown link false-positive[text](url) link syntax có [] nhưng KHÔNG phải cite. Parser yêu cầu content trong [] PHẢI là số nguyên thuần (^[0-9]+$).
  • Bibliography entry regex — entry line PHẢI bắt đầu chính xác [N] đầu line (không indent). Indent line bị skip vì có thể là continuation hoặc quote.
  • Backup file scope — KHÔNG include chapter-*-backup-YYYY-MM-DD.md trong verify mặc định. Backup là snapshot, không phản ánh current cite state.
  • Out of scope: scope của skill này — chỉ extract + verify, KHÔNG sửa source. Người dùng tự fix (sửa bib hoặc body) sau khi nhìn report.
  • Adapt to your project's path layout — default documents/<thesis-dir>/references/bibliography.md. Adjust scripts/args nếu layout khác (vd thesis/references/).

Related

  • documents/<thesis-dir>/references/bibliography.md — canonical bibliography (flat IEEE list)
  • documents/<thesis-dir>/references/CITATION-STYLE.md — IEEE format spec (in-text + entry templates) [recommended companion doc]
  • .claude/rules/skill-conventions.md — skill structure mandate
  • .claude/rules/thesis-content-standard.md C3 Bibliography category — this skill closes C3 audit gap (100% inline cite utilization + orphan detection)
  • Sister skill: quality/thesis-figure-curation/SKILL.md (closes C7 Diagram + figure rendering category)

What ships with it: 6 files

20.7 KB alongside SKILL.md, 3 of them executable

data/

scripts/

Gives 0 of the 12 instructions most quality gates skills give in ~1.3k tokens

Counted across 1,524 of the 2,830 authors here whose files we hold, read 2026-09-06

  • Read full output and check exit codein 45 of 1524, across 40 files
  • Verify output confirms the claimin 44 of 1524, across 39 files
  • Identify the command that proves the claimin 43 of 1524, across 39 files
  • Execute the full verification commandin 36 of 1524, across 30 files
  • Produce a verification reportin 34 of 1524, across 18 files
  • Review git diff changesin 30 of 1524, across 16 files
  • Fix build failures immediatelyin 29 of 1524, across 9 files
  • Group findings by severityin 28 of 1524
  • State claim only with evidencein 27 of 1524, across 22 files
  • Verify regression tests with red-green cyclein 26 of 1524, across 22 files
  • Run the full test suitein 26 of 1524, across 25 files
  • Run test suite with coveragein 25 of 1524, across 10 files

Said here and by no other author read

  • Extract citation keys from thesis chapter markdown files
  • Verify citations against the canonical bibliography file
  • Report matched citations and orphan references
  • Save verification results to a JSON cache file
  • Exclude citations found within code blocks
  • Ignore markdown link syntax during extraction

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.