Skip to content
Back to skills

External Peer Review

ASecurity

Obtain independent external frontier-class peer reviews of a manuscript PDF; synthesize convergent findings into one verdict.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 19, 2026
research

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned October 19, 2026

npx -y skills add MinhHaDuong/ImperialDragonHarness --skill external-peer-review --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of External Peer Review?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for External Peer Review
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/minhhaduong-external-peer-review/badge)](https://www.skillsdirectory.com/skills/minhhaduong-external-peer-review)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: external-peer-review
description: "Obtain independent external frontier-class peer reviews of a manuscript PDF; synthesize convergent findings into one verdict."
user-invocable: true
disable-model-invocation: false
argument-hint: "<pdf-path> [reviewer needs or personas]"
---

# External peer review

Obtain actual independent reviews and present a cross-reviewer synthesis.
This complements the simulated panel in `/review-pr-prose`.

1. **Locate the manuscript.** Resolve the PDF path and any requested review
   questions. If needed, use the project's build procedure to obtain the PDF.
2. **Describe the reviewers needed.** Request independent advisors with the
   relevant expertise (deep judgment; pick per the `route` skill).
   Prefer reviewers decorrelated from the producer and from one another.
   Personas such as a skeptical expert and an attentive student can diversify
   questions; they do not establish model independence by themselves.
3. **Let the runtime discover and route.** The runtime determines available
   reviewers and the route: llama.cpp on Padmé, OpenRouter, local agents, or
   agents on another host. No gateway, fixed roster, model ID, or transport is
   required. Honor any user constraints on where the manuscript may be sent.
   If no suitable reviewer is available, report that limitation without
   presenting a simulated review as an external one.
4. **Smoke-test one review, sequential-blocking.** Confirm that the selected
   reviewer receives the complete manuscript and returns a non-empty review.
   Use the runtime's available mechanism. The bundled OpenRouter helper is
   optional; read `references/openrouter.md` only when that route is selected.
5. **Collect the remaining reviews, parallel-background when supported.**
   Independent reviewers can work concurrently. Record completed and missing
   reviews, reviewer identities, route, persona, and input coverage in durable
   artifacts. Failed requests remain missing rather than implicit approval.
6. **Synthesize every completed review.** Present consensus verdict, convergent
   concerns weighted by independent support, and sharp individual catches.
   Preserve dissent and attribute findings to the reviewers that supplied them.

Reviews are advisory input for the author. This skill produces review artifacts
and does not publish or merge anything.

Files in this skill

  • SKILL.md2.3 KB
  • peer_review.py17.7 KB
  • references/openrouter.md4.5 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.