arxiv-source-first-paper-ingest skillA
arxiv-source-first-paper-ingest is agent-read markdown (skill) from terrylica/cc-skills: Read an arXiv paper faithfully by fetching the authors' own LaTeX source instead of OCR-ing the PDF. Extracts every display equation with its label, environment and enclosing proposition, renders each to MathML, and proves usability by compiling each formula with a real TeX engine. Use when you need a paper's formulas exactly — to replicate a method, implement a loss function or custom objective, verify a quoted number, or check a claim against what the authors actually wrote. TRIGGERS - read th.
Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.
What the file says
# arXiv source-first paper ingest > **Self-Evolving skill.** If a step here fails or the tooling drifts, fix this file in the same > change. The Post-Execution Reflection at the bottom says what to check. **The whole idea in one line: for an arXiv paper, the PDF is a rendering of something you can get losslessly, so do not read the rendering.** `arxiv.org/e-print/<id>` serves the authors' own LaTeX. It carries the exact formulas, their `\label`s, and — critically — which formulas are _stated propositions_ rather than intermediate steps inside a proof. None of that survives into a PDF, so no vision model can recover it. Measured on one real paper, OCR of the PDF reached **0.958 mean token similarity** to the source and **only 49 % of its formulas compiled**, against 98 % for the source. Full numbers, including a systematic model defect that similarity cannot see, in [`../../references/OCR-VERSUS-AUTHOR-LATEX-GROUND-TRUTH.md`](../../references/OCR-VERSUS-AUTHOR-LATEX-GROUND-TRUTH.md). --- ## Step 1 — fetch the source, not the PDF Identify yourself; arXiv asks for it and rate-limits anonymous bulk access. ```bash ID=2605.00501 UA="your-project/1.0 (mailto:you@example.com)" …
Read the whole file at its exact version.
How to install
mdr add terrylica/cc-skills/arxiv-source-first-paper-ingest@git:20260731.1bd7c64mdr add terrylica/cc-skills/arxiv-source-first-paper-ingest@sha256:6bec80e59461fc80Pin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.
[](https://markdownregistry.com/a/art_ic7eo5qlpbxnqd6d)
1 badge views in 30 days
Versions
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20260731.1bd7c64 latest | 2026-07-31 | 1bd7c64 | 5,832 B | A | view · diff |
| git:20260731.4670486 | 2026-07-31 | 4670486 | 5,833 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (5832 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
terrylica/cc-skills · 74 stars · license MIT · pushed 2026-09-24 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_ic7eo5qlpbxnqd6d GET https://markdownregistry.com/api/v1/resolve?ref=terrylica/cc-skills/arxiv-source-first-paper-ingest GET https://markdownregistry.com/api/v1/blob/6bec80e59461fc80523969c6a0baaf87061528a0e37c0bb992dceeaed15b8369
Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.