integrity-verification-agent

integrity-verification-agent is an agent for Claude Code from AlterLab-IEU/AlterLab-Academic-Skills. It costs 194 tokens per session (7,039 once invoked), scanned A, original, MIT.

An academic fact-checking agent that checks whether references exist, whether their details are correct, and whether a paper's claims match the cited sources. It also checks data and originality.

In plain words
What is it for?
Use it to verify citations and research claims before peer review and after revisions.
Why use it?
It helps catch made-up references, incorrect citation details, and claims that a source does not actually support.

Agent for Claude Code

Written for Claude Code: allowed-tools in frontmatter.

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is uv run python skills/core/alterlab-citation-verifier/scripts/verify_citations.py \.

Part of the alterlab-core plugin — 10 skills, 4 commands, 34 agents shipped together

Good fit Use it to verify citations and research claims before peer review and after revisions.

Compare 6 agents from other repositories ↓
Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo
git clone --depth 1 https://github.com/AlterLab-IEU/AlterLab-Academic-Skills
agentmods
npx agentmods add agents/alterlab-ieu/alterlab-academic-skills/integrity_verification_agent

Made for: Claude Code.

Or install alterlab-core, the plugin that ships this one along with the rest of its 10 skills, 4 commands, 34 agents.

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.

agentmods badge for integrity-verification-agent

README.md
[![agentmods](https://agentmods.dev/badge/agents/alterlab-ieu/alterlab-academic-skills/integrity_verification_agent.svg)](https://agentmods.dev/agents/alterlab-ieu/alterlab-academic-skills/integrity_verification_agent)
Your own site
<a href="https://agentmods.dev/agents/alterlab-ieu/alterlab-academic-skills/integrity_verification_agent"><img src="https://agentmods.dev/badge/agents/alterlab-ieu/alterlab-academic-skills/integrity_verification_agent.svg" alt="Measured on agentmods" height="20"></a>
Per session 194 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 7,039 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00194 $0.07039
Opus 5 $0.00097 $0.03520
Sonnet 5 $0.00039 $0.01408
Haiku 4.5 $0.00019 $0.00704

Measured 4d ago against content hash 2a1e788475bc, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

integrity-verification-agent 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 4d 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.

skills/core/alterlab-research-pipeline/agents/integrity_verification_agent.md · 513 lines

How it starts

The opening of the file, as written. The whole thing — 513 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Integrity Verification Agent — Academic Integrity Verification Gatekeeper

Verification Tooling (read first)

This agent is the executable backbone of the integrity gate. It does not verify references or claims from model memory — it shells to two deterministic scripts and only falls back to WebSearch when those scripts (or the network) are unavailable:

Tool Purpose Used in
skills/core/alterlab-citation-verifier/scripts/verify_citations.py Reference existence + metadata check: resolves against Crossref / OpenAlex / Semantic Scholar / arXiv, title+author Levenshtein >= 0.70, DOI/arXiv-ID resolution, Retraction Watch flag. Detects TF (NOT_FOUND), PAC (corrupted metadata), IH (identifier hijacking). Phase A
skills/core/alterlab-citation-verifier/scripts/claim_faithfulness.py Claim faithfulness check: retrieves the cited source and compares the paper's claim against the source's actual content, mapping to the verdict taxonomy. Detects SH (Semantic Hallucination — real source, unsupported/contradicted claim) and the Frankenstein "real paper, wrong claim" pattern. Phase E
# Existence + metadata (Phase A) — batch a .bib/.txt file or DOI/arXiv list (preferred),
# or pipe a single inline reference via stdin ('-'):
uv run python skills/core/alterlab-citation-verifier/scripts/verify_citations.py \
    references.txt --format freeform --mailto <contact-email> --threshold 0.70 \
    --out integrity_existence.json
echo "<full reference string>" | \
  uv run python skills/core/alterlab-citation-verifier/scripts/verify_citations.py -
#   add --offline when there is no network (emits 'unverified', never a silent pass)

# Claim faithfulness (Phase E) — single (claim, DOI) pair or batch JSON of pairs:
uv run python skills/core/alterlab-citation-verifier/scripts/claim_faithfulness.py \
    --claim "<exact claim text from paper>" --doi <doi> --json
uv run python skills/core/alterlab-citation-verifier/scripts/claim_faithfulness.py \
    --input claim_pairs.json --tier heuristic --json   # add --tier llm for LLM-judge

Read the full file on GitHub · 513 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.

  1. 4d ago First seen · 513 lines · 194 tokens per session scan A 2a1e788475bc

Subscribe to this mod's changes

integrity-verification-agent is an agent published in the GitHub repository AlterLab-IEU/AlterLab-Academic-Skills (66 stars, last pushed 3d ago), licensed MIT. It adds 194 tokens to every session and 7,039 once invoked, about $0.0010 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 2026-09-03.