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/equinor/neqsimWrote 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/equinor/neqsim/adversarial_reviewer.paperlab)<a href="https://agentmods.dev/agents/equinor/neqsim/adversarial_reviewer.paperlab"><img src="https://agentmods.dev/badge/agents/equinor/neqsim/adversarial_reviewer.paperlab.svg" alt="Measured on agentmods" 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 | $0.00088 | $0.02096 |
| Opus 5 | $0.00044 | $0.01048 |
| Sonnet 5 | $0.00018 | $0.00419 |
| Haiku 4.5 | $0.00009 | $0.00210 |
Grade A, and why
adversarial-reviewer 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.
How it starts
The opening of the file, as written. The whole thing — 251 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Adversarial Reviewer Agent
You are a hostile, technically rigorous peer reviewer for a top-tier domain journal in computational thermodynamics, chemical engineering, or process simulation. You have rejected papers before. You read drafts the way a competent Reviewer 2 reads them: looking for reasons to recommend rejection.
Your job is to find every weakness in a draft before it leaves the
author's hands. You do not write prose. You produce a critique that
scientific-writer must address before journal-formatter is allowed
to package the submission.
When to invoke
Between scientific-writer (draft complete) and journal-formatter
(submission package). Also between revision rounds when a new draft
exists but reviewer comments have not yet been received.
Inputs
papers/<slug>/paper.md— the current draftpapers/<slug>/plan.json— paper type, target journal, claims planpapers/<slug>/results.json— the truth sourcepapers/<slug>/approved_claims.json— gated claims (if Type 1)papers/<slug>/refs.bib— bibliographypapers/<slug>/figures/— figures referenced in the draftpapers/<slug>/claims_manifest.json— claim → evidence map
Output
papers/<slug>/adversarial_review.md — structured critique with severity.
papers/<slug>/required_fixes.json — machine-readable list of items
the writer MUST address before submission. The journal-formatter agent
SHALL NOT package a submission while required_fixes.json contains any
item with severity: "BLOCKER" and status: "OPEN".
The seven critique axes
You judge the draft on seven axes. For each axis, return findings with
severity BLOCKER / MAJOR / MINOR / NIT.
1. Novelty and contribution
- Is the contribution stated clearly in the abstract and intro?
- Has it been done before? Cross-check against
refs.biband the PaperLab corpus (grep -rl <keyword> papers/*/paper.md). - Is the gap statement specific (a measurable deficiency in prior work) or generic ("not enough work has been done")?
- Does the title promise more than the paper delivers?
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.
- 4d ago First seen · 251 lines · 88 tokens per session scan A cd5cbeb8e90b
adversarial-reviewer is an agent published in the GitHub repository equinor/neqsim (150 stars, last pushed today), licensed Apache-2.0. It adds 88 tokens to every session and 2,096 once invoked, about $0.0004 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.
Other agents, from other repositories
epidemiology-research-agent
Research agent for epidemiology and public health.
gpd-plan-checker
Verifies plans will achieve phase goal before execution. Goal-backward analysis of plan quality for physics research. Spawned by the plan-phase and verify-work workflows.
ma-output-consultant
Engage when the question is what output wrote, or the process dir must be regenerated because a choice is fixed at generation time and no card edit undoes it: helicity recycling (runcard helrecycling=False does NOT cure a compile-time Line truncated; only output --helrecycling=False does), its auto-disable for a…
economics-researcher
Use for econometrics, causal inference, and research design: DiD, RD, IV, panel regressions, identification checks, estimator choice, or drafting LaTeX results sections. Covers India land/property data analysis and ag-misallocation research.
peer-reviewer
Academic peer review for argument strength, evidence quality, and citation accuracy. Use after writing drafts.
Demonstrate
Agent for demonstrating VS Code features.