What is paper-verify-webchat?

A workflow for checking whether citations and sources in a document are valid, using a chat and web research. It extracts references, fetches each source, judges the evidence, and calculates a score with a fixed rubric.

Use it to extract citations from a document, check papers or webpages, compare claims with source passages, and produce an evidence-based citation score.

It helps reveal dead links, unsupported claims, and weak source matches instead of treating every citation as reliable.

paper-verify-webchat — Skill for Claude Code and Codex by nolainjin

Install

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.

Any agent

npx skills add nolainjin/paper-verify --skill paper-verify-webchat

Clone the repo

git clone --depth 1 https://github.com/nolainjin/paper-verify

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/nolainjin/paper-verify/paper-verify-webchat/github.svg)](https://agentmods.dev/skills/nolainjin/paper-verify/paper-verify-webchat)

Your own site

<a href="https://agentmods.dev/skills/nolainjin/paper-verify/paper-verify-webchat"><img src="https://agentmods.dev/badge/skills/nolainjin/paper-verify/paper-verify-webchat/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/nolainjin/paper-verify/paper-verify-webchat"><img src="https://agentmods.dev/badge/skills/nolainjin/paper-verify/paper-verify-webchat.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 70 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,036 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00070	$0.01036
Opus 5	$0.00035	$0.00518
Sonnet 5	$0.00014	$0.00207
Haiku 4.5	$0.00007	$0.00104

Measured 10d ago against content hash e72cc29f653a, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

paper-verify-webchat 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 10d 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.

integrations/skills/paper-verify-webchat/SKILL.md · 87 lines

How it starts

The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Paper Verify (web chat)

Citation verification from a chat. Deterministic parts (extraction, rubric scoring, report) run as bundled code; fetching and judging use your web tools. The rubric is exactly the CLI's (paperverify/score.py).

Requirements

This skill folder bundles the stdlib-only paperverify package — no pip install needed. Run all commands from this skill's directory (the folder containing this SKILL.md), so python -m paperverify resolves.
Code execution unavailable? Follow fallback-prompt.md (bundled) inline instead — same protocol, model-computed scores, clearly lower precision.

Workflow

Get the document. Save the user's text/upload to a file, e.g. /tmp/doc.md.
Extract (code):
```
python -m paperverify /tmp/doc.md --extract-only > /tmp/citations.json
```
Show the user a short table (id, type, ref, line) and the count.
Fetch + judge (web tools). For each citation object: open ref with web fetch (search the title if it fails). Then build one evidence item:
- citation: the object from step 2, verbatim.
- fetched: status (HTTP-ish: 200 ok / 404 dead / 0 unreachable), title, abstract (the relevant passage, ≤1500 chars), url_final, authors (list, if visible), year (int or null), source: "http" (or "none" when unreachable — then also set error), soft_404_suspect: true for error/placeholder pages that return 200.
- judgements: exactly one — {"judge": "webchat:claude", "verdict": Match|Partial|Mismatch|Uncertain|Inaccessible, "reason": "one line quoting the source"}. Unreachable source ⇒ Inaccessible; never guess (prefer Uncertain).
Assemble /tmp/evidence.json:
```
{"source_file": "doc.md", "level": "L2", "citations": [ ...evidence items... ]}
```
(Shape reference: bundled examples/evidence-sample.json.)

Score + report (code):

python -m paperverify --from-evidence /tmp/evidence.json --json --out /tmp/pv

Read the full file on GitHub · 87 lines

Changes

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.

10d ago First seen · 87 lines · 70 tokens per session scan A e72cc29f653a

Subscribe to this mod's changes

paper-verify-webchat is a skill published in the GitHub repository nolainjin/paper-verify (5 stars, last pushed yesterday), licensed MIT. It adds 70 tokens to every session and 1,036 once invoked, about $0.0003 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 2026-08-31.

Citation Documents LLM Research arxiv doi fact-checking mcp

academic-research

Nested swiss-knife reference for academic literature work — find papers, fetch full-text PDFs, trace citations, write LaTeX manuscripts. First action for any "get me this paper" request: python3 /scripts/fetchpaper.py — walks arXiv → Unpaywall → Europe PMC → CORE → in-house publisher-page extraction…

Lingtai-AI/lingtai · 180 tokens

cvpr-iccv-eccv

CVPR / ICCV / ECCV paper formatting — activate when the user wants to submit to CVPR, ICCV, or ECCV, follow their template, or fix format issues for these computer vision conferences.

nanoAgentTeam/research-claw · 52 tokens

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.

nanoAgentTeam/research-claw · 62 tokens

paper-verify-webchat