Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/tdupu-check-citations
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: documents, security, writing, research
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/check-citations
What's inside
- Skill name
check-citations- Description
- Read-only citation audit for .tex files. Scans \cite{} calls, validates each key against .bib file(s), flags non-pinpoint references (missing page number AND missing mathematical atom). Emits check-citations-report.{json,md} under <scratch-dir>/citations/. Used as the citation sub-skill of check-latex. Read-only: no Edit/Write/git.
- License
- GPL-3.0
- Instructions
- 133 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
> **STATUS: PRELIMINARY (DRY-RUN ONLY).** Read-only. Writes ONLY under
> `<scratch-dir>/citations/`. No `Edit`/`Write`/`MultiEdit`/`git commit`
> /`git push`/`bd update`. This is a v0 preliminary per `he-cnat`.
> Companion write-enabled skill: `clean-citations` (proposals mode in v0).
# check-citations
Read-only citation audit skill. Scans a `.tex` file for all `\cite{}` calls,
validates each key against the project's `.bib` file(s), and flags non-pinpoint
references. Produces a structured report consumed by `clean-citations`.
## When to use
Use whenever you need to audit citation quality in a `.tex` file:
- Before pushing a notes-tier `.tex` edit (the LaTeX gate requires citation evidence)
- When `check-latex` composes you as part of its gate evidence block
- To identify broken `.bib` keys or vague citations before submitting
## Workflow
1. **Locate the .bib file(s)** — search in this order:
- Same directory as the `.tex` file
- A sibling `refs/` directory
- Record `bib_files: [], note: "no .bib found"` if neither is present
2. **Enumerate citations** — scan the `.tex` file for all `\cite{...}` patterns:
- `\cite{key}`, `\cite[optional]{key}`, `\cite{key1,key2,...}`
- Split multi-key citations; record each key as a separate entry
- Capture: `key`, `tex_line`, `optional_arg` (or null), `context` (surrounding text excerpt)
3. **Check bib_match** — for each cited key, check whether the key appears as
`@...{key,` in any discovered `.bib` file. Record `bib_match: true` or `false`.
4. **Detect non-pinpoint citations** — a citation is **pinpoint** if its optional
argument contains at least one of:
- A **page reference**: matches `\bp\.?\s*\d`, `\bpp\.?\s*\d`, or `\bpage\b` followed by a number
- A **mathematical atom**: any of `Theorem`, `Proposition`, `Lemma`,
`Definition`, `Remark`, `Corollary`, `Example`, `Conjecture` (case-insensitive)
A citation with no optional argument, or whose optional argument contains
neither a page reference nor a mathematical atom, is flagged `non_pinpoint: true`.
Otherwise record `non_pinpoint: false`.
5. **Write the report** — create `<scratch-dir>/citations/`:
```
check-citations-report.json — JSON array of citation objects
check-citations-report.md — human-readable Markdown table
```
Use the current bead ID from context as `<bead>`. If no bead is active, use
a slug like `check-citations-<yyyy-mm-dd>`.
**JSON schema per entry:**
```json
{
"key": "smith1984",
"bib_match": true,
"non_pinpoint": false,
"tex_line": 47,
"optional_arg": "Theorem 3.1",
"context": "...see \\cite[Theorem 3.1]{smith1984} for the proof..."
}
```
**Markdown table format:**
```markdown
| Key | bib_match | non_pinpoint | Line | Optional arg |
|-----|-----------|--------------|------|--------------|
| smith1984 | ✓ | ✗ | 47 | Theorem 3.1 |
```
6. **Empty case** — if the `.tex` file has no `\cite{}` calls, write reports
with `citations: []` and a note `"no citations found in <filename>"`.
Do NOT raise an error.
## Report output contract
- The JSON report at `check-citations-report.json` has top-level structure:
```json
{
"tex_file": "<absolute path>",
"bib_files": ["<path>", ...],
"note": "<note if no bib found>",
"citations": [...]
}
```
- The Markdown report summarizes the same data as a table plus a brief header.
- Neither report file modifies any `.tex` or `.bib` file.
## Read-only constraint
This skill is **entirely read-only**:
- Use `Read`, `Bash` (read-only commands: `grep`, `find`, `cat`, `ls`), and
`Write` ONLY for the two report files under `<scratch-dir>/citations/`.
- Never call `Edit`, `MultiEdit`, or any write-to-source-file pattern.
- Never run `git commit`, `git push`, `git add`, or `git checkout`.
- Never call `bd update` or any bead-mutation command.
## Cross-reference
`clean-citations` is the companion write-enabled skill. It reads
`check-citations-report.json` produced by this skill and emits proposed
citation corrections (proposals-mode in v0; in-place writes in v1).
Run `check-citations` first, then pass its report path to `clean-citations`.
## Examples
**Broken citation detected:**
```
tex: \cite{smith1984} bib: no "smith1984" entry
→ {key: "smith1984", bib_match: false, non_pinpoint: true}
```
**Non-pinpoint citation flagged:**
```
tex: \cite{jones2020} (no optional arg)
bib: "jones2020" entry exists
→ {key: "jones2020", bib_match: true, non_pinpoint: true}
```From tdupu/agent-skills-public, GPL-3.0.
More from tdupu/agent-skills-public
- 0.570score
assemble-recording-series
tdupu
Assemble one lecture or talk video from a folder of raw takes, whatever the capture tool wrote - flat screen-capture files, or per-session folders with the…
SkillAFreshness: AClaude Code +9 - 0.570score
check-paragraph
tdupu
Judge ONE paragraph of prose (LaTeX allowed) as its stated reader and return a fixed six-line verdict — topic sentence, one idea, reader can follow, claim…
SkillAFreshness: AClaude Code +9 - 0.565score
check-documentation-policy
tdupu
Audit mathcity documentation against POLICY-documentation.md.
SkillAFreshness: AClaude Code +9 - 0.565score
behavior-verify
tdupu
Use when an edit to a skill, an instruction file, or a prompt has to be shown to have changed the agent behavior it was meant to fix.
SkillAFreshness: AClaude Code +9 - 0.565score
check-grant-format
tdupu
Audit an NSF proposal folder against Grants/GRANT-FORMAT.md (page limits, fonts, margins, required headings, no URLs in the description, prior-support items…
SkillAFreshness: AClaude Code +9 - 0.565score
check-grant-style
tdupu
Audit an NSF Project Description and Summary against Grants/GRANT-STYLE.md (every project has a plan and names what is new; results in prose with citations…
SkillAFreshness: AClaude Code +9 - 0.565score
check-latex
tdupu
Produce the evidence block the human adjudicator needs to approve or reject a notes.tex (or any notes-tier .tex) change before push/merge.
SkillAFreshness: AClaude Code +9 - 0.565score
check-latex-hygiene
tdupu
Audit LaTeX beads, branches, or .tex diffs against the LaTeX Subdomain Policy (mathcity/subdomains/latex/POLICY.md, LX-rules).
SkillAFreshness: AClaude Code +9 - 0.565score
check-lmfdb-hygiene
tdupu
Audit an LMFDB-type bead, diff, experiment proposal, or live type against the LMFDB Subdomain Policy (mathcity/subdomains/lmfdb/POLICY.md) — labels valid…
SkillAFreshness: AClaude Code +9 - 0.564score
behavior-cause
tdupu
Use when a reproduced agent behavior has to be traced to the artifacts whose text produced it.
SkillAFreshness: AClaude Code +9 - 0.561score
behavior-repro
tdupu
Use as step 1 when a complaint about an agent's behavior has to become evidence, before any cause is diagnosed or any artifact edited.
SkillAFreshness: AClaude Code +9 - 0.561score
adjust-behavior
tdupu
Start here when an agent did something undesired and you suspect the instructions it read — a skill, a project instruction file, a prompt template — rather…
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 14, 2026 (14 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- GPL-3.0
- Repository
- github.com/tdupu/agent-skills-public
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from tdupu
- 0.570score
assemble-recording-series
tdupu
Assemble one lecture or talk video from a folder of raw takes, whatever the capture tool wrote - flat screen-capture files, or per-session folders with the…
SkillAFreshness: AClaude Code +9 - 0.570score
check-paragraph
tdupu
Judge ONE paragraph of prose (LaTeX allowed) as its stated reader and return a fixed six-line verdict — topic sentence, one idea, reader can follow, claim…
SkillAFreshness: AClaude Code +9 - 0.565score
check-grant-format
tdupu
Audit an NSF proposal folder against Grants/GRANT-FORMAT.md (page limits, fonts, margins, required headings, no URLs in the description, prior-support items…
SkillAFreshness: AClaude Code +9 - 0.565score
behavior-verify
tdupu
Use when an edit to a skill, an instruction file, or a prompt has to be shown to have changed the agent behavior it was meant to fix.
SkillAFreshness: AClaude Code +9 - 0.565score
check-documentation-policy
tdupu
Audit mathcity documentation against POLICY-documentation.md.
SkillAFreshness: AClaude Code +9 - 0.565score
check-grant-style
tdupu
Audit an NSF Project Description and Summary against Grants/GRANT-STYLE.md (every project has a plan and names what is new; results in prose with citations…
SkillAFreshness: AClaude Code +9
Cite this page
check-citations skill. Agent Mart. https://agentmart.dev/skills/tdupu-check-citations (last updated 2026-10-24)