Verify refs
Skill yha9806/academic-writing-toolkit/.claude/skills/verify-refs
Local-first, evidence-controlled academic writing workflows for AI agents, with bounded revision, clean-room review, and release governance.
npx -y skills add yha9806/academic-writing-toolkit --skill verify-refsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
What its author says it does
Copied from the file, not written here
Use when checking BibTeX reference records for missing fields, malformed identifiers, duplicate keys, or metadata mismatches before submission.
SKILL.md
1.5 KB, 294 tokens by cl100k_base, as published. Nobody here has run it
/verify-refs — Reference Authenticity Check
Purpose
Check reference records for missing required fields, duplicate keys, malformed DOI values, malformed arXiv identifiers, and invalid URLs. The default mode is offline and deterministic.
Trigger Words
This skill activates on: verify refs, verify references, reference check, /verify-refs.
Workflow
- Identify the target Markdown or
.bibfile. If the user does not specify one, ask. - Run:
python3 scripts/verify-refs.py --bib {path} --json - Report issues by entry key and severity.
- For explicit metadata verification, run:
python3 scripts/verify-refs.py --bib {path} --json --online - Use CrossRef for DOI metadata, Semantic Scholar as a secondary metadata source, and arXiv for preprint identifiers. Online checks must be explicit because they depend on network availability.
- In tests or offline review, use
--metadata-dir {dir}to read CrossRef JSON, Semantic Scholar JSON, and arXiv Atom fixtures instead of live network calls.
Constraints
- Do not import project-specific reference rules from other repositories.
- Do not auto-fix reference records without user approval.
- Keep output plain Markdown with no emoji.
- Treat CrossRef, Semantic Scholar, and arXiv as verification sources, not as citation style authorities.
Gives 0 of the 12 instructions most quality gates skills give in 294 tokens
Counted across 1,195 of the 2,094 authors here whose files we hold, read 2026-08-07
- read the output and check the exit codein 54 of 1195, across 14 files
- verify requirements using a line-by-line checklistin 53 of 1195, across 12 files
- identify the verification command proving the claimin 51 of 1195, across 12 files
- run the full verification commandin 50 of 1195, across 11 files
- verify output confirms the claimin 49 of 1195, across 12 files
- check version control diff after agent delegationin 46 of 1195, across 6 files
- state claim with evidencein 44 of 1195, across 4 files
- run the test suitein 33 of 1195, across 26 files
- keep state in memory by defaultin 27 of 1195, across 6 files
- make prototype runnable with one commandin 26 of 1195, across 5 files
- produce a verification reportin 25 of 1195, across 14 files
- detect the package manager from lockfilesin 24 of 1195, across 5 files
Said here and by no other author read
- run the offline verification script
- report issues by entry key and severity
- use CrossRef for DOI metadata verification
- use Semantic Scholar for secondary metadata verification
- use arXiv for preprint identifier verification
- use local metadata fixtures for offline reviews
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.