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.
npx skills add bdongucla/math-ai-review-skills --skill paper-reviewgit clone --depth 1 https://github.com/bdongucla/math-ai-review-skillsWrote 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/skills/bdongucla/math-ai-review-skills/paper-review)<a href="https://agentmods.dev/skills/bdongucla/math-ai-review-skills/paper-review"><img src="https://agentmods.dev/badge/skills/bdongucla/math-ai-review-skills/paper-review/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/skills/bdongucla/math-ai-review-skills/paper-review"><img src="https://agentmods.dev/badge/skills/bdongucla/math-ai-review-skills/paper-review.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 | $0.00084 | $0.08695 |
| Opus 5.5 | $0.00034 | $0.03478 |
| Sonnet 5.5 | $0.00017 | $0.01739 |
| Haiku 4.5 | $0.00008 | $0.00869 |
Grade A, and why
paper-review 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 2mo 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 — 432 lines — stays where its author put it; the contents beside it link to each section on GitHub.
About This Skill This skill is derived from real-world research review practice. It applies to papers at the intersection of mathematics and AI, as well as pure mathematics, theoretical AI, and adjacent fields. The core methodology (evidence-based critique discipline, independent coordinate system, credit attribution) applies to all mathematics-type papers.
Paper Review Skill
Operational guidelines for paper review. Covers the review workflow, output templates, and detailed Referee Mode criteria.
Trigger contexts: user requests paper review, deep evaluation, or rapid screening.
Trigger phrases: paper review, review, evaluate, deep review, rapid screening
Domain surveys, field overviews, and technical intelligence analysis → use a dedicated survey/domain-research skill if your framework includes one.
Review Task Classification (Determine Before Proceeding)
Upon receiving a review task, classify the type before selecting a workflow. Do not default to the full iron-rule process.
| Type | Context | Workflow | Target Time |
|---|---|---|---|
| Submitted Paper — Deep Review | Reviewing a student's paper; formal peer review | Full iron-rule workflow (evidence-based discipline, full reading of prior work, chain-value analysis, second-order verification) | 30–60 min |
| PhD Thesis — External Examination | Acting as external examiner for another group's student | Lightweight workflow: quick read → focus on novelty + experimental sufficiency → holistic judgment of "meets degree requirements" | ≤10 min |
| Rapid Screening | Batch-reviewing papers; deciding whether to read in depth | Template A-Rapid Screening | ≤5 min |
Execution notes for PhD thesis external examination:
- No need to verify every claim individually against evidence
- No need to read prior work in full
- No need to perform second-order consistency checks
- Focus on: (1) what the core innovation is and whether the contribution is substantive; (2) whether experiments sufficiently support the conclusions; (3) publication record; (4) writing standards
- Extract PDF tables using
pdftoppm + image, not thepdftool (table data unreliable) - Complete in the main session; do not spawn a sub-agent (high disconnection risk for PDF-intensive reads)
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.
- 2mo ago First seen · 432 lines · 84 tokens per session scan A 59a0e026d95b
paper-review is a skill published in the GitHub repository bdongucla/math-ai-review-skills (11 stars, last pushed 6mo ago), licensed MIT. It adds 84 tokens to every session and 8,695 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 2026-08-30.
paper-review compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| paper-reviewbdongucla/math-ai-review-skills | Skill | 11 | 6mo ago | MIT | A | 84 | Claude Code, Codex |
| imagegenopenai/codex | Skill | 134k | today | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 99k | today | Apache-2.0 | A | 19 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | today | MIT | A | 200 | Claude Code, Codex |
| systematic-debuggingobra/superpowers | Skill | 314k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 95 | Claude Code |
| next-partial-prefetching-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 103 | Claude Code |