vulnerability-writeup

Generate source-validated vulnerability disclosure reports from findings, PoCs, and scanner output.

5.3k|765|Updated Mar 4, 2026
One-click install
npx skills add https://github.com/openai/plugins --skill vulnerability-writeup
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: vulnerability-writeup
Source: https://github.com/openai/plugins/tree/main/plugins/codex-security/skills/vulnerability-writeup
Command: npx skills add https://github.com/openai/plugins --skill vulnerability-writeup

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Turning raw vulnerability notes, PoCs, or scanner findings into accurate, reproducible disclosure reports is slow and error-prone, and unverified claims about affected versions or impact can undermine a report's credibility.

Core Features & Use Cases

  • Sceptical source validation: Inspects the exact vulnerable source revision, traces the introducing change, and verifies affected and fixed release versions before drafting.
  • Structured report format: Produces self-contained Markdown reports with Executive Summary, Background, Vulnerability Details, Exploitability Analysis, Proof of Concept, Remediation, and Summary sections.
  • Campaign workflow: Deduplicates multiple findings by root cause and assigns one drafting sub-agent per vulnerability for disclosure campaigns.
  • Use Case: Given a Codex Security scan finding and the target repository, produce a polished disclosure report with verified affected versions, a real PoC, and a plain-English remediation plan.

Quick Start

Use the vulnerability-writeup skill to turn these vulnerability notes and the checked-out source tree into a self-contained disclosure report.

Frequently Asked Questions about vulnerability-writeup

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a vulnerability disclosure report from scanner findings?

Provide the raw finding, PoC artefacts, and the target source tree; the skill validates each claim against the exact vulnerable revision before drafting. It produces a structured Markdown report covering impact, exploitability, proof of concept, and remediation.

How to verify affected versions for a vulnerability report?

The skill traces the vulnerable code through file history and blame to find the introducing change, then inspects release tags and branches to confirm the first affected and fixed releases. Unverified version ranges are flagged rather than claimed.

Can I write up multiple vulnerabilities in one disclosure campaign?

Yes, the campaign workflow deduplicates findings by root cause and source path, then assigns one drafting sub-agent per distinct vulnerability. The main agent independently reviews each returned draft against the pinned source before acceptance.

Does the vulnerability report require a working proof of concept?

No, a PoC is included only when real artefacts exist or can be safely developed within the stated authorisation. If execution was not possible, the report labels expected results as unobserved rather than fabricating output.

What happens if the source code contradicts the reported vulnerability?

The skill stops presenting the finding as a vulnerability and instead explains the contradiction and remaining evidence. It never manufactures a persuasive disclosure for a false positive.