What is coarse-write?

A document-generation step for a research-paper review pipeline. It reads verified review data, writes a Markdown review, and converts that review to a PDF.

It is for producing a formatted paper review in reviews/slug-review.md and reviews/slug-review.pdf.

It turns already checked classification, overview, and comment files into the final review documents without redoing the earlier analysis.

coarse-write — Skill for Claude Code and Codex by Felipe-SO: A document-generation step for a research-paper review pipeline.

Install

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.

agentmods

npx agentmods add skills/felipe-so/coarse-ink-claude-code/coarse-write

Any agent

npx skills add Felipe-SO/coarse-ink-claude-code --skill coarse-write

Clone the repo

git clone --depth 1 https://github.com/Felipe-SO/coarse-ink-claude-code

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/felipe-so/coarse-ink-claude-code/coarse-write.svg)](https://agentmods.dev/skills/felipe-so/coarse-ink-claude-code/coarse-write)

Your own site

<a href="https://agentmods.dev/skills/felipe-so/coarse-ink-claude-code/coarse-write"><img src="https://agentmods.dev/badge/skills/felipe-so/coarse-ink-claude-code/coarse-write.svg" alt="Measured on agentmods" height="20"></a>

Per session 40 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,049 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00040	$0.01049
Opus 5	$0.00020	$0.00524
Sonnet 5	$0.00008	$0.00210
Haiku 4.5	$0.00004	$0.00105

Measured 4d ago against content hash 261d34d30f0b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

coarse-write 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.

.claude/skills/coarse-write/SKILL.md · 144 lines

How it starts

The opening of the file, as written. The whole thing — 144 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/coarse-write — Write Final Review

Usage: /coarse-write <slug>

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

CLASSIFICATION = .coarse_cache/<slug>_classification.json
OVERVIEW_JSON = .coarse_cache/<slug>_overview.json
COMMENTS_VERIFIED = .coarse_cache/<slug>_comments_verified.json
OUTPUT = reviews/<slug>-review.md
OUTPUT_PDF = reviews/<slug>-review.pdf

Re-read CLASSIFICATION, OVERVIEW_JSON, and COMMENTS_VERIFIED before writing.

Step 14 — Write Output

Write the final review to OUTPUT using the Write tool. Use this exact format:

# {title}

**Date**: {today's date as MM/DD/YYYY}
**Domain**: {domain}
**Taxonomy**: {taxonomy}
**Filter**: Active comments

---

## Overall Feedback

Here are some overall reactions to the document.

{assessment paragraph if non-empty}

**{Issue 1 title}**

{Issue 1 body}

**{Issue 2 title}**

{Issue 2 body}

[... all overview issues ...]

**Recommendation**: {recommendation text}

**Key revision targets**:
1. {target 1}
2. {target 2}
[... only if recommendation is not "accept" ...]

**Status**: [Pending]

---

## Detailed Comments ({N})

### 1. {comment 1 title}

**Status**: [Pending]

**Quote**:
> {verbatim quote — prefix every line with "> " for multi-line quotes}

**Feedback**:
{feedback text}

---

### 2. {comment 2 title}
[...]

After writing, confirm the file exists with the Read tool and print the path.

See reference_review.md in the workspace root for a gold-standard example showing the expected depth, tone, and format.

Step 15 — Convert to PDF (PyMuPDF)

Run this Python one-liner to convert OUTPUT to OUTPUT_PDF:

python -c "
import fitz, markdown, pathlib, sys
sys.stdout.reconfigure(encoding='utf-8')
md = pathlib.Path('OUTPUT').read_text(encoding='utf-8')
body = markdown.markdown(md, extensions=['extra'])
html = '''<!DOCTYPE html><html><head><meta charset=\"utf-8\"><style>
body{font-family:Georgia,serif;font-size:11pt;line-height:1.55;color:#1a1a1a}
h1{font-size:15pt;line-height:1.3;margin-bottom:6pt}
h2{font-size:13pt;margin-top:18pt;border-bottom:.5pt solid #bbb;padding-bottom:2pt}
h3{font-size:11pt;margin-top:14pt}
blockquote{border-left:2pt solid #999;margin:8pt 0;padding:2pt 10pt;color:#444;font-style:italic;background:#f8f8f8}
hr{border:none;border-top:.5pt solid #ddd;margin:16pt 0}
p{margin:6pt 0}
strong{font-weight:bold}
</style></head><body>''' + body + '''</body></html>'''
story = fitz.Story(html=html)
writer = fitz.DocumentWriter('OUTPUT_PDF')
mediabox = fitz.paper_rect('a4')
where = mediabox + (40, 50, -40, -50)
more = 1
while more:
    dev = writer.begin_page(mediabox)
    more, _ = story.place(where)
    story.draw(dev)
    writer.end_page()
writer.close()
print('PDF written: OUTPUT_PDF')
"

Read the full file on GitHub · 144 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

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 · 144 lines · 40 tokens per session scan A 261d34d30f0b

Subscribe to this mod's changes

coarse-write is a skill published in the GitHub repository Felipe-SO/coarse-ink-claude-code (2 stars, last pushed 4mo ago), licensed MIT. It adds 40 tokens to every session and 1,049 once invoked, about $0.0002 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 2026-08-31.

Documents PDF

pydicom

Use pydicom to read, inspect, write, transform, and safely preflight local DICOM datasets and pixel data. Applies to DICOM metadata, transfer syntaxes, compression plugins, frames, private elements, JSON, and bounded de-identification review.

K-Dense-AI/scientific-agent-skills · 56 tokens

foundry-hosted-agent-validation

Step-by-step process for validating a Python Foundry hosted agent sample (under python/samples/04-hosting/foundry-hosted-agents/) end to end — running it locally (native runtime and azd ai agent run) and after deploying it to an Azure AI Foundry project with azd. Use this when asked to validate a hosted agent sample.

microsoft/agent-framework · 82 tokens

coarse-write