# Coarse Classify

> Coarse pipeline steps 3-6 — classify paper domain/taxonomy, produce domain calibration, extract contributions, and search literature. Writes classification, calibration, contribution, and literature cache files.

- **Type:** Skill
- **Install:** `agentstack add skill-felipe-so-coarse-ink-claude-code-coarse-classify`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [Felipe-SO](https://agentstack.voostack.com/s/felipe-so)
- **Installs:** 0
- **Category:** [Search](https://agentstack.voostack.com/c/search)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [Felipe-SO](https://github.com/Felipe-SO)
- **Source:** https://github.com/Felipe-SO/coarse-ink-claude-code/tree/main/.claude/skills/coarse-classify

## Install

```sh
agentstack add skill-felipe-so-coarse-ink-claude-code-coarse-classify
```

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

## About

# /coarse-classify — Classify, Calibrate, Contribute, Literature

**Usage**: `/coarse-classify `

Argument (`$ARGUMENTS`) is the paper slug (e.g. `my-paper`). All paths are relative
to the workspace root `d:/Dropbox/Research/Coarse Reviewer/`.

## Path setup

- `EXTRACTED` = `.coarse_cache/_extracted.md`
- `SECTIONS_JSON` = `.coarse_cache/_sections.json`
- `CLASSIFICATION` = `.coarse_cache/_classification.json`
- `CALIBRATION` = `.coarse_cache/_calibration.json`
- `CONTRIBUTION` = `.coarse_cache/_contribution.json`
- `LITERATURE` = `.coarse_cache/_literature.txt`

Read `SECTIONS_JSON` and the first 2000 chars of `EXTRACTED` before starting.

---

## Step 3 — Classify

Determine:
- **title**: exact paper title from the first page
- **domain**: e.g. `social_sciences/economics`, `computer_science/machine_learning`,
  `statistics/causal_inference`, `natural_sciences/biology`
- **taxonomy**: e.g. `academic/research_paper`, `academic/review_paper`,
  `academic/working_paper`
- **abstract**: the paper's abstract text
- **math_sections**: list of section indices (0-based) whose `math_content` should
  be `true`. A section needs math verification if it contains ANY of: proofs
  (formal or informal), theorem/lemma/proposition/corollary statements with
  arguments, formal definitions or assumptions, algebraic manipulations, estimator
  definitions, asymptotic expressions.

**Save to `CLASSIFICATION`** using the Write tool:
```json
{
  "title": "...",
  "domain": "...",
  "taxonomy": "...",
  "abstract": "...",
  "math_sections": [0, 2, 4]
}
```

---

## Step 4 — Domain Calibration

Using the paper's title, domain, abstract, and section list, produce a
domain-specific review calibration.

You are an expert academic reviewer. For each field below, provide 3-5 concise
items tailored to this paper's specific domain and methodology:
1. **methodology_concerns**: The key methodological concerns for this type of paper
2. **assumption_red_flags**: Assumptions that commonly fail in this domain
3. **what_not_to_check**: What is irrelevant for this paper type
4. **evaluation_standards**: What a top-tier journal in this field expects

**Save to `CALIBRATION`** using the Write tool:
```json
{
  "methodology_concerns": ["...", "..."],
  "assumption_red_flags": ["...", "..."],
  "what_not_to_check": ["...", "..."],
  "evaluation_standards": ["...", "..."]
}
```

---

## Step 5 — Contribution Extraction

Read the abstract, introduction, and conclusion sections from `EXTRACTED`.

You are an expert academic reader. Extract the paper's stated contributions,
key mathematical objects, and author defenses. Your task is READING COMPREHENSION
— report what the paper SAYS, not your assessment of it.

- **main_claims**: Quote or closely paraphrase each contribution the paper
  explicitly states. Include the specific mathematical result, not generic descriptions.
- **key_objects**: The central mathematical objects/quantities and what the paper
  claims about each.
- **stated_limitations**: Any limitations the authors explicitly acknowledge.
- **author_defenses**: Objections the authors anticipate and address, including
  the section/remark where the defense appears.
- **methodology_type**: The paper's approach in one sentence.

**Save to `CONTRIBUTION`** using the Write tool:
```json
{
  "main_claims": ["...", "..."],
  "key_objects": ["...", "..."],
  "stated_limitations": ["...", "..."],
  "author_defenses": ["...", "..."],
  "methodology_type": "..."
}
```

This is a hard constraint for all later steps. Any review comment that
contradicts the stated claims without a concrete counterexample must be dropped.

---

## Step 6 — Literature Context (Optional)

If WebSearch is available, search for 3-5 closely related papers:
- Query: `"[paper title]" related work`
- Query: `[core method] [domain] [year range]`

Format the results as a numbered list:
```
1. **Title** (Authors, Year) — brief relevance note
```

**Save to `LITERATURE`** using the Write tool (plain text, numbered list).
If no results found, write an empty file. This file is read by the overview
and section review steps.

---

Done. Next: `/coarse-overview `

## Source & license

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

- **Author:** [Felipe-SO](https://github.com/Felipe-SO)
- **Source:** [Felipe-SO/coarse-ink-claude-code](https://github.com/Felipe-SO/coarse-ink-claude-code)
- **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-felipe-so-coarse-ink-claude-code-coarse-classify
- Seller: https://agentstack.voostack.com/s/felipe-so
- 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%.