{
  "markdown": "# Deep Review\n\nDeep Review is an experimental skill for rigorous review of papers and technical artifacts.\n\nIt asks a short calibration interview, reads the artifact carefully, decomposes the review into specialist passes, checks claims against sources and computations where possible, self-critiques its own findings, and returns a severity-ranked review.\n\nThe goal is simple: find the errors a generic chat-with-file pass is likely to miss.\n\nDeep Review can also maintain optional review memory — user-level taste in `~/.deep-review/` and project-level history in `.deep-review/` — so future reviews can pick up the user's preferred standards, recurring concerns, and review taste. It only writes memory after explicit user approval.\n\n## What It Is Good For\n\n- pre-submission paper review\n- theorem or proof stress tests\n- empirical / methods audits\n- literature positioning checks\n- technical memo review\n- AI-generated draft review\n\n## Status\n\nDeep Review is now live in Anthropic's\n[Claude Community marketplace](https://github.com/anthropics/claude-plugins-community).\nIt is still early, and I am learning where it is most useful.\n\nIf you want to try it on a real paper or artifact, install the skill in your coding agent.\n\n## Install\n\n### Claude Code\n\nInstall Deep Review from the Claude Community marketplace:\n\n```text\n/plugin marketplace add anthropics/claude-plugins-community\n/plugin install deep-review@claude-community\n/reload-plugins\n```\n\nThe repository-hosted marketplace is also available as a fallback:\n\n```text\n/plugin marketplace add dipakkrishnan/deep-review-skill\n/plugin install deep-review@deep-review-marketplace\n/reload-plugins\n```\n\nOr copy the skill folder into a project directly:\n\n```bash\nmkdir -p .claude/skills\ncp -R /path/to/deep-review-skill/skills/deep-review .claude/skills/deep-review\n```\n\nThen ask Claude Code:\n\n```text\nUse $deep-review to review this paper for theorem correctness and literature positioning.\n```\n\n### Claude.ai\n\nDownload [`deep-review.zip`](https://github.com/dipakkrishnan/deep-review-skill/releases/latest/download/deep-review.zip) (always points at the latest release), then in Claude.ai go to **Customize → Skills → + → Upload a skill** and drop the zip in.\n\nTo build the zip yourself from a checkout:\n\n```bash\nscripts/package-skill.sh\n# produces dist/deep-review.zip\n```\n\n### Codex\n\nCodex does not yet expose a marketplace-style installer for third-party plugins. Copy the skill into a Codex skills directory and restart Codex:\n\n```bash\ncp -R skills/deep-review \"$CODEX_HOME/skills/deep-review\"\n```\n\n## Example Prompts\n\n```text\nUse $deep-review to review this PDF as a skeptical economics referee. Focus on model correctness, calibration, and missing literature.\n```\n\n```text\nUse $deep-review to stress-test the proof of Theorem 2 and identify assumptions that do not hold.\n```\n\n```text\nUse $deep-review to review this technical memo. Skip copy edits and focus on unsupported claims.\n```\n\n## What To Expect\n\nDeep Review will usually ask a few calibration questions first:\n\n- What is your relationship to the artifact?\n- What kind of criticism is most useful?\n- Which parts worry you most?\n- What should it avoid spending time on?\n\nThen it produces a markdown review with:\n\n- overall assessment\n- critical issues\n- significant concerns\n- minor issues\n- questions for the authors\n\nOptionally, Deep Review can save local memory for future runs:\n\n- standing review preferences\n- target venues or standards\n- recurring issue patterns\n- lightweight history of reviewed artifacts\n\n## Requirements\n\nThe skill itself is instructions and references.\n\nThe helper scripts require Python. PDF extraction uses PyMuPDF:\n\n```bash\npip install pymupdf\n```\n\n## Privacy\n\nDeep Review does not collect telemetry or send data to any third-party service controlled by this plugin. See [PRIVACY.md](PRIVACY.md).\n\n## Notes\n\nDeep Review is not a replacement for expert human judgment. It is a way to make review cheaper, sharper, and more adversarial before you ask humans to spend scarce attention on the artifact.\n",
  "bytes": 4048,
  "sha": "e3c01b151bb3511b3bc2b15cd3b3b12f4104de2cf0abf7f305346b9a4be71772",
  "repo_slug": "dipakkrishnan/deep-review-skill",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_dipakkrishnan_deep_review_skill_deep_rev_da1142c1/readme"
}