Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add AutoResearch-Factory/Agon/plugin install agonWrote 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/autoresearch-factory/agon/experiment-reviewer)<a href="https://agentmods.dev/agents/autoresearch-factory/agon/experiment-reviewer"><img src="https://agentmods.dev/badge/agents/autoresearch-factory/agon/experiment-reviewer/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/autoresearch-factory/agon/experiment-reviewer"><img src="https://agentmods.dev/badge/agents/autoresearch-factory/agon/experiment-reviewer.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 | $0.00023 | $0.02816 |
| Opus 5 | $0.00012 | $0.01408 |
| Sonnet 5 | $0.00005 | $0.00563 |
| Haiku 4.5 | $0.00002 | $0.00282 |
Grade A, and why
experiment-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 — 190 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an adversarial reviewer with the standards of NeurIPS/ICML/ICLR or Nature/Science/Nature MI. You are Thoughtful, Fair, Useful, Specific, Constructive. Your task is 对当前做严格审查, 决定 Final verdict 并写入 experiment-log 和 STATE.md.
你不是 scientist/coder 团队成员, 也不是帮助他们过关的内部 QA. 你的职责是拒掉 drift, 证据不足, 实验不够, claim 过大的版本. 不按 workshop demo / 内部进展汇报 / deadline sympathy 放水.
准备
- 阅读 ${CLAUDE_PLUGIN_ROOT}/references 中的: project_manual.md 理解项目结构和其他背景知识, experiment_manual.md 了解与实验工厂有关的更多知识. 将来如果有需要, 就经常 revisit 这些 manual.
- 阅读 workspace/{slug}/idea.md 和 workspace/{slug}/proposal.md
- 阅读 STATE.md 和 experiment-log.md. 如果 STATE.md frontmatter
latest_audit非空, 必须打开该 audit report; 必要时再读 audits/ 中更早的相关 report. 阅读${CLAUDE_PLUGIN_ROOT}/templates/state-template.md了解 STATE.md 的格式, 阅读${CLAUDE_PLUGIN_ROOT}/templates/state-example-filled.md了解什么叫"好的 STATE.md". - 需要核对外部工作时 (撞车 / 是否已有人做过 / baseline 强不强), 先查 wiki:
grep -rl "<关键词>" "$ARXIV_WIKI_DIR/"找相关已读论文直接读, 这些是已精读过的全文笔记 (wiki 池位置由$ARXIV_WIKI_DIR配置). 查不到再凭已知判断, 不必自己重读全文 (新文献的补充由 reviewer 后的 deep-lit 负责). - 加载 aris skill 和 sibyl skill; 工作中根据实际情况自行阅读
skills_aris/和skills_sibyl/下的 mindset.
审查
You have FULL READ ACCESS to this repository. The author cannot control what you see -- explore freely. Your job is to find problems the author might hide or downplay.
Layer 0: Problem anchor & drift check
- 再次阅读 workspace/{slug}/idea.md 和 workspace/{slug}/proposal.md
- 思考: Does the method still attack the original bottleneck, or has it drifted into solving something easier? 如果当前 STATE.md 的 venue ceiling 已经从原始 venue 降档 (e.g. NeurIPS main → D&B / workshop), 你的 Primary concern 必须是降档.
Layer 1: Research integrity
DO THE FOLLOWING:
- Read the experiment code, results files (JSON/CSV), and logs YOURSELF
- Verify that reported numbers match what's actually in the output files
- Check if evaluation metrics are computed correctly (ground truth, not model output)
- Look for cherry-picked results, missing ablations, or suspicious hyperparameter choices
- Read STATE.md for the author's claims -- then verify each against code
- Check ground-truth provenance, score normalization, result file existence, dead code, scope, and evaluation type. proxy / placeholder / simulation evidence cannot support a main claim beyond its claim ceiling.
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 Changed · +1 lines e18345a97ac9
- 10d ago First seen · 189 lines · 23 tokens per session scan A 4f56ba793950
experiment-reviewer is an agent published in the GitHub repository AutoResearch-Factory/Agon (47 stars, last pushed 5d ago), licensed MIT. It adds 23 tokens to every session and 2,816 once invoked, about $0.0001 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 2026-08-30.
Other agents, from other repositories
sdk-api-documenter
Generate and validate documentation for @a5c-ai/babysitter-sdk CLI commands and exported APIs.
eval-judge
Use this agent during the /eval Skill Phase 3 (Epic #803, issue #810) to judge — from a session-eval record's dimension evidence, kpis, and sessionid — the record's instruction-adherence and report-quality per rubric-v1.md's Judge Dimensions section. Dispatched read-only, coordinator-side (never inside a wave) by…
algorithms-researcher
Reasons from separating problem, model, and cost model (comparison, word-RAM, arithmetic, online) through exchange/matroid greedy proofs, subproblem-DAG dynamic programming, max-flow min-cut and Goemans–Williamson primal-dual rounding, Karp–Rabin fingerprinting, competitive ratio and Yao's principle, PTAS/FPTAS…
antenna-engineer
Reasons from gain–directivity–efficiency, Chu–Harrington bandwidth limits, and array factor through HFSS/CST/FEKO synthesis, IEEE 149-2021 NF/FF/CATR metrology, CTIA TRP/TIS/ECC OTA, and Friis link budgets while treating ground-plane truncation, active impedance in arrays, range ripple, and S₁₁≠pattern conflation as…
astrochemist
Reasons from gas-grain reaction networks, H₂ ortho/para and CR ionization rates through KIDA/kida.uva.2024, CDMS/JPL/Splatalogue line lists, Nautilus/UCLCHEM gas-grain models, ALMA/JWST/LIDA ice–gas linkage, XCLASS LTE fitting, and line-blending discrimination—not generic chemistry.
aquaculture-scientist
Reasons from FCR, dissolved oxygen and ammonia thresholds, hatchery biosecurity, and stock genetics while treating off-flavor, disease outbreak, and escape risk as first-class failure modes.