# Paper Claim Extractor

> Use when reading an ML paper, abstract, README, or result section and the task is to extract claims, assumptions, experiments, limitations, or implementation requirements.

- **Type:** Skill
- **Install:** `agentstack add skill-maximussthegreat-ml-researcher-os-paper-claim-extractor`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [maximussthegreat](https://agentstack.voostack.com/s/maximussthegreat)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [maximussthegreat](https://github.com/maximussthegreat)
- **Source:** https://github.com/maximussthegreat/ml-researcher-os/tree/main/skills/paper-claim-extractor

## Install

```sh
agentstack add skill-maximussthegreat-ml-researcher-os-paper-claim-extractor
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Paper Claim Extractor

Use this skill before implementing, reproducing, or summarizing an ML paper.

## Goal

Convert paper text into a structured research brief that makes the agent slow down before coding.

## Required output

Produce these sections:

1. `Core claim`
2. `Evidence table`
3. `Assumptions`
4. `Datasets and splits`
5. `Metrics`
6. `Baselines`
7. `Ablations`
8. `Implementation requirements`
9. `Limitations`
10. `Questions before reproduction`

## Evidence table format

| Claim | Evidence in text | Evidence type | Confidence | Reproduction note |
| --- | --- | --- | --- | --- |

Evidence type must be one of:

- `reported_result`
- `method_description`
- `ablation`
- `theory`
- `author_claim`
- `missing`

Confidence must be one of:

- `high`
- `medium`
- `low`

## Rules

- Do not treat the paper's conclusion as ground truth.
- Do not invent datasets, hyperparameters, or baselines.
- If a detail is missing, write `not specified`.
- Separate what the paper reports from what you infer.
- Flag any claim that depends on a single dataset, a weak baseline, or missing ablations.

## Final check

End with:

```text
Ready to plan experiments: yes/no
Blocking missing details:
- ...
```

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [maximussthegreat](https://github.com/maximussthegreat)
- **Source:** [maximussthegreat/ml-researcher-os](https://github.com/maximussthegreat/ml-researcher-os)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-maximussthegreat-ml-researcher-os-paper-claim-extractor
- Seller: https://agentstack.voostack.com/s/maximussthegreat
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.