---
name: cfp-cybersecurity-review
description: Review and improve an author-written cybersecurity conference CFP draft using Black Hat-informed criteria. Use for titles, abstracts, outlines, supporting answers, or complete proposals aimed at research and professional security conferences. Do not use to originate a Black Hat submission from scratch.
---

# Cybersecurity CFP review

Coach the author before submission. Be skeptical about evidence and generous about useful revision. The score describes what the draft demonstrates; it is not an acceptance probability.

Read [references/rubric.md](references/rubric.md) before scoring.

## Intake

Accept an incomplete draft, but identify fields you cannot assess. Seek: target event and track, title, abstract, detailed outline, contribution versus prior work, audience takeaways, relevant speaker access, tools/demos/findings, commercial relationship, prior or planned publication, and vulnerability disclosure status.

Never treat an unanswered publication, relationship, or disclosure question as clean.

## Review

1. Read for the claim, method, evidence, result, limits, and audience value. Distinguish a dramatic target from a novel technique.
2. Verify novelty with current authoritative sources when browsing is available. Name the search and sources briefly. Otherwise label novelty unverified.
3. Score all ten rubric dimensions out of 100. Missing material earns no assumed credit.
4. Check whether the outline contains the work rather than generic section names, whether a tool supports research or is the whole pitch, and whether commercial framing displaces evidence.
5. Rank three to five revisions by likely substantive effect. Quote the author's line when diagnosing it and offer a refinement that preserves the author's actual claim.

For Black Hat, follow the current event policy: help edit or refine author-written material and review prior art; do not generate a submission from scratch. Never invent findings, CVEs, affected versions, measurements, citations, credentials, disclosure status, or release promises.

## Response

Lead with `score/100` and a plain-language readiness band. Then provide the dimension table, prioritized revisions, open questions, and specific strengths worth preserving. State any novelty-verification limit. When the draft is already strong, focus on venue fit, audience reach, current affected versions, realistic deliverables, and sales framing instead of manufacturing a fatal flaw.

End by reminding the author to check the current CFP and that selection also depends on competition, program balance, and expert judgment.


---

# Included reference: rubric.md

# Cybersecurity conference rubric

| Dimension | Max | Full-credit evidence |
|---|---:|---|
| Novelty and differentiation | 20 | Specific contribution, closest prior work, credible difference, and current relevance |
| Technical depth and outline | 17 | Followable method with targets, versions, evidence, findings, and limitations |
| Practical value | 10 | Attendees can test, reproduce, detect, build, use, or decide something concrete |
| Evidence and deliverable | 14 | Real finding, tool, demo, dataset, code, or repeatable result described specifically |
| Vendor neutrality | 10 | Research-led framing, disclosed relationships, alternatives, and bounded product claims |
| Speaker credibility | 8 | Verifiable first-hand access or relevant work; prior fame is not required |
| Claims realism | 5 | Scope and impact match the evidence; limitations are explicit |
| Audience draw | 6 | Clear reason this audience will choose the session |
| Disclosure maturity | 5 | Notification, dates, embargo, patch status, and safe presentation when applicable |
| Track and audience fit | 5 | The proposal belongs in the selected program and format |

Bands: 85–100 strong on the page; 65–84 solid with gaps; 45–64 borderline; below 45 not ready. These are coaching bands, not historical acceptance probabilities.

Apply these judgment rules:

- A topic is not a contribution. Score the specific new result.
- A famous target increases draw, not novelty.
- Earlier work is neutral when the proposal clearly adds substantial material. An unchanged recorded talk is a major timeliness problem.
- Dense jargon is not technical depth. The method must be followable.
- Useful defensive work can offset incremental novelty when the output is real and broadly usable.
- Marketing words are editing findings; a commercial relationship becomes a scoring problem when the talk's logic depends on the product.
- A tool can strengthen a research talk. If feature demonstration is the center of gravity, advise checking the event's tool program.
- First-time speakers can receive strong credibility scores when their access and work are clear.
- Do not penalize an inapplicable disclosure category; award it neutrally and explain why.