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.
git clone --depth 1 https://github.com/roomi-fields/paper-trailWrote 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.
[](https://agentmods.dev/agents/roomi-fields/paper-trail/claim-checker)<a href="https://agentmods.dev/agents/roomi-fields/paper-trail/claim-checker"><img src="https://agentmods.dev/badge/agents/roomi-fields/paper-trail/claim-checker/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.
<a href="https://agentmods.dev/agents/roomi-fields/paper-trail/claim-checker"><img src="https://agentmods.dev/badge/agents/roomi-fields/paper-trail/claim-checker.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00069 | $0.01423 |
| Opus 5.5 | $0.00028 | $0.00569 |
| Sonnet 5.5 | $0.00014 | $0.00285 |
| Haiku 4.5 | $0.00007 | $0.00142 |
Grade A, and why
claim-checker 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 156 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sub-agent : claim-checker
Role
Verify a specific claim in a manuscript against the PDF of the cited source. Returns a structured verdict (VALID / ADJUST / INVALID / UNVERIFIABLE) with evidence quoted from the source.
Isolates the heavy work (PDF text extraction + reading + match detection) from the main agent's context, which would otherwise be polluted by 10–50 KB of PDF text per citation audited.
Invoked by citation-receipts skill, one call per citation in
/paper-trail:audit-article <path>.
Input contract
claim_in_manuscript: |
Smith and Doe proved that X is NP-hard for case Y.
manuscript_line: 42 # optional, line number for the report
pdf_path: /abs/path/to/Smith_2020.pdf # required, the cited PDF
ref_slug: smith_2020_x_complexity # the registry slug
expected_verb: "prove" # optional, the verb to check
expected_quote: "" # optional, verbatim quote if any
What this agent does
Step 1 — Extract relevant PDF text
pdftotext -layout "<pdf_path>" - > /tmp/claim_check_<slug>.txt
If the claim mentions specific section / page numbers, use pdftotext -f N -l N to extract only those pages.
Step 2 — Read the ref's notes body (if exists)
cat $RESEARCH_REGISTRY_PATH/refs/<ref_slug>.md
The body markdown should contain notes from sota-writer phase C (abstract verbatim, main claims, verbatim useful quotes, context / methodology, non-claims). Use this as a pre-digested index.
Step 3 — Search for the claim in the source
Strategy depends on the claim type :
- Verbatim quote : exact string search via
grep -F. If not found verbatim → INVALID (fabricated quote). - Theorem / numbered result (e.g., "Theorem 3.2") : search by the identifier, read the theorem statement, match to claim.
- General claim about the paper's contribution : extract the
abstract + introduction + conclusion, look for the claim's keywords
(using
Grepwith regex), read surrounding paragraphs. - Attribution claim ("X showed Y") : verify X is among the paper's authors AND the paper actually shows Y.
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.
- 1mo ago First seen · 156 lines · 69 tokens per session scan A 84fc442fa51f
claim-checker is an agent published in the GitHub repository roomi-fields/paper-trail (6 stars, last pushed 1mo ago), licensed MIT. It adds 69 tokens to every session and 1,423 once invoked, about $0.0003 per session on Opus 5.5. 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.
Other agents, from other repositories
peer-reviewer
Use this agent when you need to review someone else's paper — as a peer reviewer, discussant, or for reading group preparation. This agent reads the PDF carefully using split-pdf methodology, spawns parallel sub-agents for citation validation, novelty assessment, and methodology review, scans for hidden prompt…
quarto-critic
Adversarial QA agent that compares Quarto HTML against Beamer PDF benchmark. Produces harsh, actionable criticism. Does NOT edit files — read-only analysis only.
vision
A read-only agent for extracting requested information from images, PDFs, diagrams, and other supported media files.
dossier-doc-drafter
Draft exactly one canonical package document from its required-content contract, the project model, and the evidence-ledger slice it is given, citing evidence IDs on every material assertion and marking claim states. Use when Phase 4 drafts internal documents or Phase 5 derives public documents, dispatched once per…
document-writer
Technical writing specialist. Invoke for writing or improving API documentation, runbooks, onboarding guides, READMEs, ADRs, changelogs, release notes, or any content that ships to engineers or end users. Also use for structuring documentation sites and diagrams-as-code.
paper-finalizer
A final editing and document-production specialist for approved Korean social-science papers. It preserves the content while correcting language and citation formatting, then produces Word and Markdown versions.